Eruptive chronology of the Acoculco caldera complex – A resurgent caldera in the eastern Trans-Mexican Volcanic Belt (México)

Martha Gabriela Gómez-Vasconcelos

Journal of South American Earth Sciences Volume 98, March 2020, 102412 https://doi.org/10.1016/j.jsames.2019.102412 1 Eruptive chronology of the Acoculco caldera complex – a resurgent 2 caldera in the eastern Trans-Mexican Volcanic Belt (México) 3 by 4 5 6 7 Denis Ramón Avellán1, José Luis Macías2, Paul W. Layer3, Giovanni Sosa-Ceballos2, Martha Gabriela Gómez-Vasconcelos4, Guillermo Cisneros-Máximo2, Juan Manuel Sánchez-Núñez5, Joan Marti6, Felipe García-Tenorio2, Héctor López-Loera7, Antonio Pola8, and Jeff Benowitz3 8 1CONACYT 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 – Instituto de Geofísica, Universidad Nacional Autónoma de México, Antigua Carretera a Pátzcuaro 8701, 58190 Morelia, Michoacán 2Instituto de Geofísica, Universidad Nacional Autónoma de México, Antigua Carretera a Pátzcuaro 8701, 58190 Morelia, Michoacán 3College of Natural Science, Mathematics and Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK 99775 4CONACYT – Instituto de Investigaciones en Ciencias de la Tierra, Universidad Michoacana de San Nicolás de Hidalgo, Francisco J. Mujica s/n, Felicitas del Río, 58040, Morelia, Michoacán 5Instituto Politécnico Nacional-CIIEMAD, Miguel Othón de Mendizábal s/n. Col. La Escalera, C.P. 07320 Del. Gustavo A. Madero, Ciudad de México, México 6Instituto de Ciencias de la Tierra Jaume Almera, CSIC, LLuis Sole Sabaris, s/n, 08028 Barcelona, Spain 7División de Geociencias Aplicadas, Instituto Potosino de Investigación Científica y Tecnológica A.C.; Camino a la Presa San José 2055, Lomas 4a Sección, C-P. 78216, San Luis Potosí S.L.P. 8Escuela Nacional de Estudios Superiores, Universidad Nacional Autónoma de México, Antigua Carretera a Pátzcuaro 8701, 58190 Morelia, Michoacán 32 Submitted to Journal of South American Earth Sciences *Corresponding author e-mail: denisavellan@gmail.com 33 34 October 29, 2019 1 36 ABSTRACT 37 The Acoculco caldera complex (ACC) is located in the eastern part of the Trans-Mexican 38 Volcanic Belt in the northern part of the State of Puebla. The complex sits at the intersection of 39 two regional fault systems with NE-SW and NW-SE orientations. Acoculco ACC was built upon 40 atop Cretaceous limestones, the Zacatán basaltic plateau of unknown age, early Miocene domes 41 (~12.7-10.98 Ma), and Pliocene lava domes (~3.9-3 Ma). Detailed field mapping and stratigraphy 42 studies complemented assisted by 40Ar/39Ar and 14C dating allowed to divide the ACC volcanic 43 succession into identify 30 volcanic units of the complex. Based on all these data and previous 44 studies the Acoculco ACC eruptive chronology was grouped in four eruptive phases: syn-caldera, 45 early post-caldera, late post-caldera, and extra-caldera. Inception of the ACC volcanism began 46 around 2.7 Ma with the dispersion of an andesitic ignimbrite followed by the collapse of the 47 magma chamber roof as attested by the presence of a lithic breccia in isolated parts of the caldera 48 rim. The collapse produced a 18×16 km caldera depression which was partly filled by the 49 ignimbrite (total volume of ~127 km3) followed by the establishment of an intracaldera lake of 50 unknown total extension. Early post-caldera collapse activity (2.6-2.1 Ma) was restricted within 51 the caldera producing 27 km3 of lava flows and domes dominantly of basaltic trachyandesite to 52 basaltic composition. Late post-caldera collapse activity (2.0-<0.016 Ma) migrated dominantly to 53 the caldera rim and periphery emplacing 90 km3 of magma as rhyolitic domes, lavas, scoria 54 cones, and two younger ignimbrites. The 1.2 Ma Encimadas ignimbrite (26 km3) was generated 55 vented throughon the eastern margin of the caldera and dispersed to the northeast, and the 0.6-0.8 56 Ma Tecoloquillo ignimbrite and dome (11 km3) was producederupted from at the southwestern 57 margin of the ka calderascoria cones. The most recent eruption of this phase was vented close to 58 the southeastern caldera rim producing the Cuatzitzingo (<16,710 ± 50 years BP) scoria cone. 59 Extra-caldera activity (2.4-0.19 Ma) of the Apan–Tezontepec volcanic field produced scoria 2 60 cones and lava flows of basaltic trachyandesite to basaltic andesite composition that are 61 interbedded with the products of the caldera complex. 62 Aeromagnetic data further constraint the edge of the caldera rim and revealed is consistent with 63 the presence of at least four intrusive bodies at depths of >2 1 km hosted in the Cretaceous 64 limestones. These bodies might represent a series of horizontal mafic intrusions located at 65 different depths (183 ka) that besides heating and mixing with the Acoculco magma reservoirs, 66 provides the energy that maintain active the Acoculco geothermal system. 67 Keywords: Geology; geochronology; geothermal energy; Acoculco caldera; Puebla 68 INTRODUCTION 69 Silicic collapse calderas are volcanic depressions resulting from the subsidence of the 70 magma chamber roof caused by the rapid withdrawal of magma during the course of an explosive 71 eruption (Lipman, 2000; Martí et al., 2008; Geyer and Martí, 2014). The formation of a collapse 72 caldera is still an enigmatic geological phenomenon because of the structural complexity 73 involved in such type of volcanic eruptions. In addition, silicic caldera eruptions represent a high 74 risk due to the large amount of magma and eruption rates involved (Costa and Martí, 2016). The 75 resulting caldera depression may represent the site of important ore deposits and high enthalpy 76 geothermal reservoirs. This makes the study of caldera systems of great interest for modern 77 societies; as they represent a high risk but also a significant source of economic resources. In 78 many active volcanic areas, such as Iceland, New Zealand, Costa Rica, Japan, Indonesia, among 79 several others, an important part of their government energy requirements is covered by the 80 exploitation of geothermal resources associated with collapse caldera systems. The current global 81 economic crisis has increased the interest for exploring and exploiting this type of geothermal 82 resources, so a great number of research programs have been initiated all around the world to 3 83 identify and quantify this potential energy resources. In México, geothermal energy has been 84 used since 1959, when a program led by the Comisión Federal de Electricidad (CFE) started to 85 explore and exploit some of these potential energy resources, most of them located in calderas of 86 the Trans-Mexican Volcanic Belt (TMVB) (Hiriart et al., 2011). A successful exploitation of the 87 geothermal resources requires a detailed exploration of the hydrothermal reservoir, which needs 88 to be conducted using a precise combination of geology, geophysics and modeling. In the 89 particular case of geothermal fields installed in collapse calderas, it is very important to 90 determine the mechanism that formed the caldera and the post-collapse internal structure (Bibby 91 et al., 1995; Di Napoli et al., 2011; Molina et al., 2014; Afanasyev et al., 2015). to better 92 understand fluid paths and the location of reservoirs. It is crucial to determine the exact structural 93 limits of the caldera (e.g., Molina et al., 2014), the stratigraphy and age of the caldera-forming 94 deposits and the distribution of thermal anomalies. On this respect, the Acoculco caldera located 95 in the eastern TMVB (Fig. 1) represents an excellent case scenario to study the internal structure 96 of the edifice, its eruptive chronology, and surface manifestations of geothermal activity with 97 respect to fractures and faults. Acoculco has been considered for years a dry-hot rock reservoir by 98 CFE (Lorenzo-Pulido et al., 2010) and is the site of an on-going European-Mexican effort to 99 develop geothermal energy from non-conventional sources called GeMex Project (Calcagno et 100 al., 2018). 101 In this contribution, we analyzed previous studies combined with new fieldwork to 102 present a simplified volcanological map and a new composite stratigraphic column of Acoculco. 103 The results are assisted by 24 new 104 units that combined with previous information set a refined evolution model of the caldera. This 105 information is crucial to present the chronology of the caldera volcanic complex throught time. In 106 order to gain new insights on the caldera structure we produced airborne aeromagnetic models of 40Ar/39Ar ages and one 4 14C radiometric date of the volcanic 107 the area, analyzed the morphostructural distribution of the volcanic units, and their correlation 108 with subsurface units in geothermal wells. With Base on thisthese new information, we propose 109 that the late Pleistocene shallow intrusions are still the heat source of geothermal activity beneath 110 Acoculco. 111 GEOLOGICAL SETTING 112 The Acoculco Caldera Complex (ACC) is located within the TMVB, a calc-alkaline 113 volcanic arc produced by the subduction of the Cocos and Rivera plates beneath the North 114 American plate at the Middle American Trench (Pardo and Suárez, 1995) (Fig. 1). The Acoculco 115 area is located in the straddles the central and eastern sectors of the TMVB (Pasquaré et al., 116 1991). Acoculco occurs ~140 km northeast of Popocatépetl volcano which defines in this region, 117 the active front of the TMVB (Siebe et al., 1995; Macías et al., 2012) (Fig. 1). Acoculco is 118 located over a 45-50 km-thick continental crust (Urrutia and Flores-Ruíz, 1996), and sits at 400 119 km from trench where the Cocos plate plunges into the mantle (Pérez-Campos et al., 2008). This 120 area is under an NW-SE oriented extensional regime, as deduced from the alignment of volcanic 121 vents, dike orientations, extension fractures and kinematics of faults at the Apan-Acoculco region 122 (García-Palomo et al., 2002; 2018). The Acoculco caldera is bounded to the west by the Apan- 123 Tezontepec volcanic field (ATVF) (~3–0.2 Ma) (García-Palomo et al., 2002; García-Tovar et al., 124 2015), to the north and to the east by Cretaceous limestones of the Sierra Madre Occidental 125 (Avellán et al., 2018), and to the south by early Miocene volcanics of the TMVB (García-Palomo 126 et al., 2002). 127 Prior to the 80´s the Acoculco geology has not been extensively studied (e.g., Ledezma- 128 Guerrero, 1987; Castro-García and Cordoba, 1994), in fact, CFE produced the first regional map 129 of the area (De la Cruz-Martinez and Castillo-Hernández, 1986) followed by other detailed 130 studies as rock dating (López-Hernández and Castillo-Hernández, 1997; López-Hernández and 5 131 Martínez, 1996). More recentLater geophysical studies attempted to understand the internal 132 structure of the caldera (Alatriste-Vilchis et al., 2005; Campos-Enriquez et al., 2003) and the 133 hydrological features conditions (Huizar-Álvarez et al., 1997) of the Apan region. The first 134 volcanological study of the region was carried out by López-Hernández et al. (2009) who 135 concluded that Acoculco was a 18-km wide caldera active from ~1.7 to ~0.2 Ma. These authors 136 considered that Acoculco was nested within the largest 32-km wide Tulancingo Caldera, which 137 was active between ~3.0 and ~2.7 Ma. Recently, Avellán et al. (2018) presented the first detailed 138 geologic map and stratigraphy of the caldera assisted supported by nine 139 authors concluded that the caldera has a semi-circular shape (18–16 km) and that was active from 140 2.7 to 0.06 Ma, thus, not corresponding with the previous timing of the caldera formation (López- 141 Hernández et al., 2009). This geology of Acoculco was used to present a simplified geologic map 142 of the caldera to describe its geochemical evolution (Sosa-Ceballos et al., 2018) and a 143 preliminary 3D model of the Acoculco subsurface structure (Calcagno et al., 2018). 144 METHODS 145 40Ar/39Ar dates. These Twenty-four whole-rock samples, phenocryst-free, were crushed, sieved and washed in 40Ar/39Ar 146 deionized water for isotopic dating (Table 1). For most samples phenocrysts-free 147 ground mass chips (whole rock) were separated for dating. For sample Ac11, both a plagioclase 148 mineral separate and a whole rock sample were separated and dated, and for Ac69, two different 149 rock types were analyzed. The samplesThe samples were irradiated in position 5C at the 150 McMaster University Nuclear Reactor in Hamilton, Canada, for 0.75 MWh. The standard mineral 151 TCR-2 with an age of 28.619 Ma (Renne et al., 2010) was used to calculate the irradiation 152 parameter, “J”. Samples Ac39 and Ac123, were irradiated with the standard mineral MMHb-1 153 with an age of 523.5. The standards were fused and, with the exception of AC-90. The standards 154 were fused and samples step-heated using a laser dating system consisting of 6W argon-ion laser 6 155 at the Geophysical Institute, University of Alaska Fairbanks, following the technique described in 156 Layer (2000) and Layer et al. (2009). For Sample Ac90, an obsidian, 7 small shards were fused, 157 with 6 of the 7 yielding enough gas to calculate a fusion age. The samples were analyzed in a 158 VG3600 mass spectrometer and the measured Ar isotopes were corrected for system blank, mass 159 discrimination and Ca, K, and Cl interference reactions, according to procedures outlined in 160 McDougall and Harrison (1999) and using the standards reported inof Renne et al. (2010). 161 System blanks were 2x10-16 mol 40Ar and 2x10-18 mol 36Ar, which are 5 to 50 times smaller than 162 fraction volumes. Mass discrimination was monitored by running calibrated air shots. 163 The structural analysis of the caldera system comprised a review of previous and new 164 structural data, after the evaluation of digital elevation models, aerial photographs, topographic 165 maps, fieldwork and a morphological evaluation of the landforms. The fault geometry was 166 characterized with field measurements and geomorphic analyses of the faults using ArcGIS on a 167 15 m-resolution digital elevation model from INEGI (Instituto Nacional de Estadística y 168 Geografía). Grid references on maps are in the WGS 1984 UTM Zone 14N projection. Rozeta 2.0 169 software was used for plotting the fault datas on stereographic projections (lower hemisphere) 170 and to perform trend-frequency analysis with rose diagrams. Stereonet 9 software was used for 171 plotting the fault planes and poles, from where the compression and tension areas were inferred 172 into the synthetic right dihedral diagram from the two pairs of conjugate fault systems, according 173 to their slip models (e.g., De Vicente et al., 1992). 174 To estimate the volume of the volcanic units we used the geological map, the Digital 175 Elevation Model (DEM), Spot-6 satellite images (1.5-m panchromatic and 6-m multispectral), 176 and the shaded relief DEM (15-m resolution) in the ArcMap 9.3 software. The difference 177 between the actual topography and the geomorphologic element of each unit was used to obtain a 7 178 z value, and to create a 3D surface geology on the shaded relief DEM and the volume with the 179 Surface Difference tool in ArcMap 9.3. 180 VOLCANIC STRATIGRAPHY 181 In this study we analyzed the regional (García-Palomo et al., 2002; López-Hérnandez et 182 al., 2009) and local geologic maps of Acoculco the ACC and its surroundings (Avellán et al., 183 2018; Sosa-Ceballos et al., 2018; Calcagno et al., 2018), and their geochronological data include 184 in them. We compiledto present a new simplified geologic map (Fig. 2) supported by 24 new 185 40Ar/39Ar 186 Acoculco units, two pre-caldera, 30 ACC, and nine extra-caldera units. The ACC units were 187 grouped subdivided in syn-caldera, early post-caldera, and late post-caldera, corresponding to 188 different phases of the caldera formation defined by their distribution, stratigraphic position, age, 189 mineral and chemical composition (Fig. 3). The chemical variations of the ACC rocks have been 190 previously documented and discussed and will not be treated in this contribution (Sosa-Ceballos 191 et al., 2018). Next, we succinctly described all units that span in age from ~2.7 Ma to <0.016 Ma 192 with their location given with respect to the caldera rim (Figs. 2 and 3). 193 Basement 194 Undifferentiated Cretaceous Limestones (Ksl) and one 14C dates (Tables 1 and 2). The maps shows includes three basement pre- 195 The oldest rocks in the region are Cretaceous limestones (Ksl) of the Sierra Madre 196 Oriental that are exposed to the northeastern, eastern, and southeastern parts of the mapped area 197 (Fig. 2). Good outcrops appear in the Tenexapa and Ajajalpan canyons, close to the towns of 198 Chignahuapan and Zacatlán. At the Chignahuapan hot springs these rocks occur as light-gray 199 parallel stratified limestones with chert concretions between stratified beds sometimes affected by 200 vertical joints fractures filled with hydrothermal minerals (e.g., calcite). Although these rocks are 201 not exposed inside the ACC, they were cut in the two geothermal CFE wells at depths of ~1,200 8 202 m in EAC-1 (López-Hernández et al., 2009), and 350-450 m in EAC-2 (Viggiano-Guerra et al., 203 2011), likely offset by a fault (Fig. 2). 204 According to these authors, Ksl was intruded by a light-gray phaneritic granite found at 205 the bottom of EAC-1 drill hole at depths from 1800 to 2000 m. The intrusion produceding a 206 methamorphic aureole of skarns (Viggiano-Guerra et al., 2011). A sample analyzed of this well 207 resulted to be an apliteic dike is made of alkali feldspar, plagioclase, quartz, amphibole, chlorite, 208 and Fe-Ti oxides. An isochron 40Ar/39Ar age of this sample yielded an age of 183 ± 36 ka (Table 209 1, isochron age; Fig. 4A), which may correspond to eithera younger intrusions beneath Acoculco 210 (e.g. dike swarms)or likely within a reset age of an older a regional plutonic body (López- 211 Hernández et al., 2009; Calcagno et al., 2018).,. whose older age could have been reset by these 212 younger intrusions. 213 Zacatlán basaltic plateau (Za) 214 Za is aA dark-gray, up to 200 m thick basaltic lava plateau dubbed Zacatlán (Za)that, 215 discordantly overliesying the Cretaceous limestones near the Chignahuapan and Zacatlán towns 216 (Fig. 2). The lava flow is aphanitic with columnar jointing and spheroidal weathering. The age of 217 this unit is unknown, however, its stratigraphic position indicates that Za is older younger than 218 the Cretaceous limestonesm.iddle Miocene. 219 Peñuela dacitic dome complex (~13 to 10 Ma) (Mv) (Pdd) 220 These rocks are exposed around southern the ACC and belong to the first beginnings of 221 regional volcanism of the eastern part of the TMVB (Fig. 2). The oldest pre-caldera rocks are the 222 Peñuela dacitic (Pdd unit) and Quexnol andesitic dome complexes exposed to the SW and SE 223 parts of the ACCAcoculco, respectively (Figs. 2 and 3). The Peñuela dome was dated with the K- 224 Ar method at 12.7 ± 0.6 Ma (García-Palomo et al., 2002). Another sample of this rock (Pdd) was 225 dated here with the 40Ar/39Ar method at 10.98 ± 0.07 Ma (Table 1, plateau age; Fig. 4B). 9 226 Pre-caldera units (~3.9 – 3.0 Ma) (Pc) 227 The Puente (Pald) and Terrerillos (Tdld) lava domes are exposed to the N-NW and SW of 228 the geologic map, respectively (Figs. 2 and 3). These rocks are light-gray andesitic to dacitic lava 229 domes with greenish-gray enclaves. Pald yielded in this work an isochron integrated age of 230 3,.6208,883 ± 220.71 kaMa (Table 1; Fig. 4C). García-Tovar et al. (2015) reported a K-Ar age of 231 3.0 ± 0.4 Ma for Tdld, which is coherent consistent with its stratigraphic position. 232 Syn-caldera unit (Acoculco andesitic ignimbrite, ~2.7 Ma) (Sc) 233 The Acoculco andesitic ignimbrite unit (Aai) corresponds to a yellow to white massive, 234 andesitic ignimbriteunit. It consists of rounded pumice and angular to sub-angular accidental lava 235 fragments (gray, pink and banded) fragments supported by a matrix of coarse to fine ash (Fig. 236 5A). Tube-banded pumice (greenish-gray to white) with alkali feldspar and amphibole 237 phenocrysts is also present. The Aai crops out in small gullies and in some places of high 238 topographic relief where it is greatly covered by younger deposits (e.g., Hbl, Srl, Atal, Fig. 2). To 239 In the southwestern and northern parts of the caldera, Aai underlies ≤40 m thick lacustrine 240 deposits (Fig. 5B). The contact between Aai and the underlying Pdd and Pald units was not 241 observed, however, accidental fragments of these units occur in the ignimbrite. At site 24, occurs 242 a massive lithic breccia (mlBr) made of heterolithologic lavas and scarce pumice (Figs. 2 and 5C- 243 D). This breccia (5-7 m thick) contains angular accidental lavas fragments set in a coarse to fine 244 ash matrix. Some lLava fragments are aphanitic (light-gray and ochre), and other porphyritic 245 (dark-gray, greenish, and ochre) composed ofwith clinopyroxene and plagioclase phenocrysts. 246 The mlBr matrix contains reddish aphanitic lithics and loose isolated crystals of plagioclase, 247 pyroxene and amphibole, disseminated pumice and silty minerals. Laterally, this mlBr grades to a 248 massive fine ash layer enriched in pumice and crystals of plagioclase and amphibole. In three 249 sites (24, 65, and 114) mlBr is interbedded between flow units of Aai (Figs. 2 and 5D). 10 250 The Aai unit was transected at the EAC-1 exploratory well at depths between 210 and 560 251 m, where it covers the Pald unit (Pre-caldera units) and is overlain by the Pedernal rhyolitic lava 252 unit (Pdl, which is a Llate post-caldera unit) (Fig. 2). Well EAC-1 shows nearly 800 m of 253 volcanic materials resting on atop a skar which wasa interpreted as the metamorphosed 254 calcareous basement (skarn) (López-Hernández et al., 2009; Viggiano-Guerra et al., 2011). These 255 authors interpreted the volcanic column from top to bottom as: Acoculco ignimbrite (0-130 m), 256 Cruz Colorada dacite (130-210 m), Alcholoya ignimbrite (210-580 m), and las Minas rhyodacite 257 (580-790 m). However, based on the new geologic map and revised stratigraphy of Acoculco, we 258 consider that the successions would correspond to our Pdl, Aai and Pald units (Figs. 2 and 3). 259 The Pedernal rhyolitc lava (Pdl) unit crops out around the CFE drill holes as a highly altered 260 vesicular rock with phenocrysts of feldspars, plagioclase, quartz and mafic unrecognizable 261 minerals together withand recrystallized lithic clasts and fully corroded, and highly vesicular lava 262 fragments. We consider that these highly altered lavas were mistaken with pumice blocks and 263 erroneously described as the Acoculco ignimbrite. In addition, we propose that their Alcholoya 264 ignimbrite (~2.7 Ma; López-Hernández et al., 2009) described between 210 and 580 m depth 265 would corresponds to our Acoculco ignimbrite Aai (Figs. 2, 3 and 5A). Unfortunately, Aai was 266 not recognized in the succession of well EAC-2 (Fig. 2). This well only shows 340 m of volcanic 267 infill (Pdl and Pald units) resting over the skarn including 200 m of the porphyritic unit 268 (Viggiano-Guerra et al., 2011) or our Pdl unit. As mentioned before, EAC-1 and EAC-2 wells are 269 500 m apart but their stratigraphy suggests that they are separated by a fault that is not 270 recognizable at surface. By averaging these thicknesses in the EAC-1 exploratory well, gullies 271 and outcrops described at surface (~470 m) we estimate a minimum Aai volume of 127 km3 (Fig. 272 2 and 3). 11 273 The AaiA pumice fragment separated from Aai was dated with the 40Ar/39Ar method in 274 plagioclase at 2,732 ±185 ka (Avellán et al., 2018). In this work, we obtained two new whole 275 rock 40Ar/39Ar dates of Aai pumice samples that yielded younger ages of 2,041 ± 38, and 2,185 ± 276 65 ka (Table 1, plateau age; Figs. 4D-E). However, Aai underlies several early-post caldera lava 277 flows as Aguila (Atal, 2.44 Ma), and Manzanito (Mtal, 2.2 Ma), and the Sayula dome (Srl, ~2.55 278 Ma) (see figures 2 and 3). Based on these stratigraphic relationships, we concluded that the age of 279 Aai must be older than 2.55 Ma for which the 40Ar/39Ar in plagioclase (2,732 ±185 ka) is the best 280 age approximation of the formation of the caldera collapse. 281 A sequence of ≥ 40 m thick lacustrine sediments (ls) is exposed inside the caldera rim in 282 the south, southwest, and northern parts. It consists of a tilted alternation of white clayed laminae, 283 and dark-gray cm-thick, volcaniclastic beds. These beds are made of rounded lava fragments set 284 in a fine-grained matrix barren of fossils. At sites 69 and 119, ls overlies the Acoculco ignimbrite 285 (Aai) (Figs. 2 and 3). 286 287 Early post-caldera units (~2.6-~2.1 Ma) (Epc) 288 Four of these units are exposed inside the caldera depression as basaltic to basaltic 289 trachyandesite lava flows (Hbl, Atal, Vtal and Mtal) that partially cover Aai (Figs. 2 and 3). They 290 are highly eroded and have asymmetric morphologies similar to flatirons with their apex towards 291 the center of the caldera developing a sub-radial exorheic drainage. These units typically appear 292 as light-gray to dark-gray, blocky lava flows with porphyritic to aphanitic textures. These lavas 293 frequently present greenish, yellowish and reddish hydrothermal alteration zones, and host 294 xenoliths of sub-rounded limestones, sandstones and fine grain granite. Two units were dated 295 with the 40Ar/39Ar method at 2,323 ± 48 ka (Vtal) and 2,199 ± 24 ka (Mtal) (Table 1, plateau age; 296 Figs. 4F-GD). Avellán et al. (2018) obtained a 40Ar-39Ar 12 whole rock age of 2,441 ± 234 ka for 297 Atal. The absolute age of Hbl is unknown; nevertheless, it underlies the Srl dome dated at ~2.55 298 Ma (see below). Other two early-postcaldera units (Srl and Atad) are exposed on the external 299 northwestern and southeastern rim of the caldera, respectively (Fig. 2). The Srl unit (2,553 ± 110 300 ka; Table 1; Fig. 4H) consists of gray to black and brown, banded obsidian lava flows with 301 holohyaline texture and partially devitrified to light-gray spherulites and lithophysae (Fig. 3). The 302 Atad unit (2,179 ± 26 ka; Table 1; Fig. 4I) is a blocky greenish to light-gray, porphyritic lava 303 flow (Table 1). For all the early post-caldera units we estimated a volume of 27 km3 (Fig. 3). 304 Late post-caldera units (~2.0 - ~0.016 Ma) (Lpc) 305 These units are represented by 11 domes (Alrd, Lrd, Amrc, Prld, Crcd, Trcd, Crd, Ard, 306 Crl, Arcd and Mrcd), 5 lava flows (Coal, Tal, Cual, Pdl and Plrld), 2 ignimbrites (Eri, Tr1l), and 307 4 scoria cones (Plc, Tlc, Clc1 and Clc2). The Alrd, Lrd, Amrc, Crcd, Trcd, Crd, Ard, Crl Arcd 308 and Mrcd dome and lava flow units occur on the caldera border and periphery (Fig. 2). These 309 structures have predominant rhyolitic compositions with coulée and asymmetrical morphology 310 delimited by very steep levées. The rocks of these units typically appear as light-gray to pinkish- 311 gray, banded to massive obsidian lavas with mottled structure given by abundant spherulites and 312 lithophysae (Fig. 3). López-Hernández et al. (2009) and García-Tovar et al. (2015) reported K-Ar 313 ages in hornblende and whole-rock of the Lrd (1,700 ± 400 ka), Crd (1,300 ± 600 ka), and Crl 314 (1,274 ± 27 ka). In this work, seven new units were dated with the 315 yielding ages of 1,870 ± 36 ka (Alrd) (Fig. 4J4E), 1,438 ± 24 ka (Amrc) (Fig. 4KG), 1,394 ± 8 ka 316 (Crcd) (Fig. 4L), 1,360 ± 15 ka (Trcd) (Fig. 6A), 1,283 ± 88 ka (Ard) (Fig. 6B4H), 998 ± 36 to 317 1,145 ± 14 ka (Arcd) (Fig. 6C-D4J) and 1,066 ± 42 ka (Mrcd) (Fig. 6E4K), that are in agreement 318 with their stratigraphic position. On the other hand, on the western edge the undifferentiated 319 Maguey unit formed of a sequence of pyroclastic surges and fall with a 40Ar-39Ar age of 1,084 ± 320 22 ka (Table 1; Fig. 6F). 13 40Ar/39Ar method (Table 1) 321 The Coal, Tal, Cual, Pdl and Prld units are situated inside the caldera partially overlying 322 the syn-caldera and early post-caldera units (Fig. 2). They appear as a stack of bedded lava flows 323 with andesitic (Coal, Tal and Cual) and rhyolitic (Pdl and Prld) compositions. The andesitic lava 324 flows (Coal, Tal and Cual units) are porphyritic, black to dark-gray in color, with reddish to 325 yellowish intense hydrothermal alteration (Fig. 3). The Coal unit was dated by Avellán et al. 326 (2018) with the 327 yielded 1,600 ± 35 ka (Table 1; Fig. 6G). In this study we dated the Tal unit with the same 328 method at 1,708 ± 54 ka (Table 1; Fig. 6H 4F). López-Hernández et al. (2009) obtained a 329 40Ar/39Ar 40Ar-39Ar method at 2,027 ± 40 ka. A dike 40Ar-39Ar age that cuts this unit age of 1,600 ± 200 ka for the Cual unit. 330 The rhyolitic lava flows (Pdl and Plrd units) are pinkish-gray to pinkish-white porphyritic 331 rocks. These units are white and highly hydrothermally altered lavas that crop out in the vicinity 332 of Pedernal and Acoculco towns (Fig. 2). The lava flows are corroded by hydrothermal fluids 333 along highly vesicular breccia structures, where phenocrysts and matrix have been replaced by 334 mineral alteration minerals. López-Hernández et al. (2009) reported two K-Ar ages for these 335 lavas of 1,600 ± 100 ka (Pdl), and 1,400 ± 200 ka (Prld). 336 The Encimadas unit (Eri) is a rhyolitic ignimbrite widely exposed on the east-northeast 337 external parts of the ACC (Fig. 2). It has a moderately dissected peneplain that mantles the 338 Zacatlán basaltic plateau and partially covers some of the post-caldera units. Eri is a welded 339 ignimbrite with several flow units that appear as massive, light-gray to white, beds. Each bed 340 consists of matrix-supported fine ash particles with abundant feldspar and quartz phenocrysts. It 341 has an approximated volume of 26 km3 (Fig. 3). López-Hernández et al. (2009) reported a 342 40Ar/39Ar 343 obtained in this study (Table 1, plateau age; Fig. 6I4I). age in sanidine of Eri at 1,300 ± 200 ka. A similar 40Ar/39Ar age of 1,278 ± 14 ka was 14 344 The Tecoloquillo ignimbrite (Tr1) is a rhyolitic ignimbrite widely exposed to the south- 345 southwest parts of the caldera (Fig. 2). South of the caldera, Tr1 partially covers some units 346 belonging to the pre-caldera, syn-caldera and post-caldera. The Tr1 unit consists of two main 347 beds; the lowermost part is massive, monolithologic, brittle and matrix-supported with highly 348 friable pumice fragments embedded in a medium to fine ash matrix. Both pumice and matrix 349 contain bipyramidal quartz and alkali feldspar phenocrysts. The upper part is massive with pink- 350 gray, corroded lava blocks, supported by a crumbly medium ash matrix. The upper part of Tr2 is 351 a rhyolitic dome made of angular light-pink to gray lava blocks. The rock is moderately vesicular, 352 fibrous and porphyritic with quartz, alkali feldspar and amphibole phenocrysts. One pumice 353 sample collected from the basal part of the Tr1 unit yielded a 40Ar/39Ar age in plagioclase of 611 354 ± 72 ka (Table 1; Figs. 6J). Other pumice sample that yielded older age of 762 ± 9 ka (Table 1; 355 Fig. 6K). López-Hernández et al. (2009) reported a 40Ar/39Ar age in sanidine of 0.8 ± 0.1 Ma for 356 this unit. 357 The last four late-post-caldera units are cinder cones (Paila, Tuliman, Cuatzitzinguito and 358 Cuatzitzingo) and associated lava flows of basaltic andesite composition. These scoria cones 359 occur above or close to the topographic caldera rim (Fig. 2). The Paila (Plc) and Tuliman (Tlc) 360 cinder cones are exposed on the southeastern and northwestern parts of the caldera rim, 361 respectively. The Paila unit directly overlies the Atad and Eri post-caldera units and Tuliman unit 362 lies discordantly on top of Srl post-caldera unit. The Cuatzitzinguito (Clc1) and Cuatzitzingo 363 (Clc2) Other scoria cones are lies on the southern flank of the Paila unit (Cuatzitzinguito and 364 Cuatzitzingo units) (Figs. 2 and 3). All these scoria cones are composed of massive poorly-sorted 365 fallout beds with dense blocks and bread-crust scoria and spatter bombs, as well as, black to dark- 366 gray blocky lava flows associated with effusive activity. The Tulimán scoria cone was dated with 367 the 40Ar/39Ar at 63 ± 9 ka (Avellán et al., 2018). A 40Ar-39Ar age of 71 ± 17 ka was obtained here 15 368 for the Paila scoria cone (Table 1; Fig. 6L). A paleosol underneath a fallout tephra of the the 369 Cuatzitzinguito scoria cone deposit yielded a 370 obtain an absolute age for the Cuatzitzingo scoria cone, however, its stratigraphic position 371 indicates that it should be youngerst than 16 ka. 372 14C age of 16,710 ± 50 BP (Table 2). We did not We estimated a total volume for all the late-post caldera units including Eri and Trl of 90 373 km3 (Fig. 3). 374 Extra-caldera units (~2.4-0.19 Ma) (ATVF) 375 NineteenTwentyEighteen scoria cones and threefour small-shield volcanoes of the ATVF 376 occur around the caldera complex (Figs. 2 and 3). Four of these scoria cones, known as 377 Amanalco (Asc; ) (2,408 ± 58 ka; Avellan et al., 2018), Huixtepec (##Hsc), Tecolote (Tsc) and 378 Apapasco (Asc) are located ca. 7 km to the southeast of the caldera border. Five scoria cones 379 called (Buenavista (##Bsc), Comal (##Csc), Calandria (##Csc), Toronjil (##Tsc), and Tezontle 380 (Tsc##)) and the Coatzetzengo small- shield-volcano are located at ca. 4 km to the northwest. 381 Three scoria cones named Moxhuite (Msc; ) (239 ± 34 ka; Avellán et al., 2018), Matlahuacala 382 (Mlcsc), and Cazares (Csc) lie discordantly on top of the Encimadas ignimbrite at ca. 6 km to the 383 east. Two small- shield volcanoes, Camelia (Callc; ) (2,033 ± 84 ka; Avellán et al., 2018) and 384 Tetelas (Tlcal,) (1,060 ± 84 ka; Avellán et al., 2018) are located at 10 km to the south of the 385 caldera border. Four Three scoria cones are aligned in a NW-SEW direction, these are Cuate 386 (##), Tecajete (##Tsc) (1,235 ± 62 ka), Blanco (Bsc; ) (1,274 ±62 ka; Avellán et al., 2018) and 387 Hermosa (Hsc##). Another, fourthree scoria cones, Coliuca (Clc; ) (188 ± 6 ka K-Ar age; García- 388 Tovar et al., 2015), Coloradoyote (##Csc), El Conejo (Csc) and Tezoyo (Tsc##), and the Coyote 389 small-shield volcano are situated at ca. 7 km to the southwest of the caldera border. 390 CHARACTERIZATION OF FAULT SYSTEMS 16 391 The ACC is in a highly tectonized area region affected by NW-SE, NE-SW and E-W 392 structures (Fig. 6A7A). The NW-SE and NE-SW structures belong to two regional fault systems.: 393 the NW-striking Taxco-San Miguel de Allende and the NE-striking Tenochtitlan-Apan. The NW- 394 striking (NNW to NW) Taxco-San Miguel de Allende fFault sSystem contains the oldest regional 395 structures. Some of tThese faults werefollow the trend of inherited from the older structures, such 396 as fold axes and Cretaceous folds and thrust faults from the Laramide orogeny in the Sierra 397 Madre Oriental (Suter, 198491; Rocha et al., 2006; Lermo et al., 2009), locatexposed just north 398 of the study area. The NW-SE system began its activity ~30 Ma with the NE-oriented Cenozoic 399 extension (Henry and Aranda-Gómez, 1992) was active in this region from middle to late 400 Miocene, creating NW-striking normal faults and NE-striking strike-slip faults synchronous with 401 regional volcanism that represent the southernmost part of the Basin and Range province in 402 México (Aranda-Gómez and McDowell, 1998; Aguirre-Díaz et al., 2005)(Andreani et al., 2008),. 403 In the ACC, the NW-SE fault system is represented by several major normal and oblique (right- 404 lateral component) faults and numerous minor fault strands with lengths between 2 and 5 km, 405 with an average azimuthstrike of 130˚ dipping mainlyboth to the NE and SW at average angles of 406 50˚, creating horst and graben-like structures (García-Palomo et al., 2002; ) (Fig. 67A). The 407 Manzanito structure, also known as the Tulancingo-Tlaxco fault system (López-Hernández et al., 408 2009), is the most representative fault pertaining to this system in the ACC (Avellán et al., 2018). 409 It has an en échelon array of normal faultsstructure, which extends for ca. 30 km that measures 410 ca. 30 km (Fig. 6B) and displacesing the 1.7 Ma Lobera rhyolithic dome by at least 145 m, and 411 the 1.07 Ma Minilla rhyolitic dome by 120 m (this study; Fig. 6B). Also, atThe most recent 412 activity of this system in the ACC occurred around 1.6 Ma, when the ~160˚-striking and ~2.5 km 413 long Colorada basaltic dikes were emplaced at the southeastern part of the caldera (Fig. 2; Coal; 414 Avellán et al., 2018). These dikes, together with parallel faults indicate a NE-oriented Pleistocene 17 415 tectonic extension that was synchronous with volcanism. But there is no evidence of recent 416 (Holocene) activity in the local NW-striking faults. 417 The NE-striking (NNE to NE) Apan-Tlaloc Fault System (García-Palomo et al., 2018), 418 Tenochtitlan-Apan Fault System (García-Palomo et al., 2002) or Tenochtitlan Shear Zone (De 419 Cserna et al., 1988), has been active since the Miocene and is still active (García-Palomo et al., 420 2002; 2018; Fig. 76). This is the most important fault system in the region, consisting of normal 421 faults with a left-lateral component with an average azimuthstrike of 040˚ dipping both to the 422 NW and SE with an average dip angle of 75˚ that present two generations of striae (~30 and 80°). 423 This system has created regional horst and graben-like structures, obeying a NW-oriented 424 extension regime that affects eastern México (this study and García-Palomo et al., 2018; ) (Fig. 425 67A). In general, this faultit system shows a good geomorphic expression represented by several 426 major faults (e.g. Apan-Tlaloc Fault and Chignahuapan Fault) and numerous minor fault strands 427 and fractures with 1-4 km individual lengths. Field exposures show prominent scarps; the ~1.34 428 Ma Mesa ChicaCanoas rhyolitic dome is displaced by the Atotonilco faultscarp (northern caldera 429 rim) by 150 m, and the ~2.2 Ma Ajolotla trachyandesitic dome is displaced by the Chignahuapan 430 fault by 200 m (this study). The most recent volcanic structures of the ACC (Plc with 71 ka, and 431 Tlc with 63 ka units) are cut by these faults (this study; Figs. 2 and 67B). 432 The NE- and NW-striking normal fault systems intersect each other (García-Palomo et al., 433 2002; Lermo et al., 2009) creating an orthogonal arrangement of grabens, half-grabens and 434 horsts. The NE-SW Rosario-Acoculco Horst (García-Palomo et al., 2002) is delimited to the west 435 by the ~235°-trendingstriking and NW-dipping Apan-Tlaloc Fault (Mooser and Ramírez, 1987; 436 Huizar-Álvarez et al., 1997) and to the east by the ~055°-trendingstriking and SE-dipping 437 Chignahuapan Fault (Avellán et al., 2018defined in this study). The ACCcoculco Graben is a 438 volcano-tectonic grabendepression inside the Rosario-Acoculco Horst, delimited by parallel 18 439 faults but with opposite dipping. As for the E-W structures, they are represented by at least 10 E- 440 W striking (~N085°) and SE-dipping normal fault strands. They can only be observed inside the 441 ACC, mainly affecting post-caldera volcanism in the central part of the ACC (this study; Figs. 2 442 and 76). 443 AEROMAGNETIC DATA 444 We produced an aeromagnetic map by reprocessing the airborne data of the Mexican 445 Geological Survey obtained in 2000 (Fig. 78A). This map shows different aeromagnetic 446 anomalies inside or outside of the ACC. The outer domains may be caused by regional and 447 topographic anomalies. Instead, inner domains are possibly associated to four positive local 448 anomalies (subdomains) and conforming a semi-circular shape of the Acoculco caldera. The first 449 subdomain is located in the central part of the ACC; it shows an NE-oriented elongated shape 450 (9.7 km long and 4.8 km wide) with magnetic intensities between -81.8 and 57.8 nT. The second 451 anomaly is located at the southeastern boundary of the ACC, ENE-oriented, 6.6 km long and 4.1 452 km wide with magnetic intensities between -135.1 and 25.7 nT. The third anomaly is located in 453 the south-central portion of the ACC, NE-oriented, 8.9 km long and 3.5 km wide with magnetic 454 intensities between -64.5 and 36.5 nT. The fourth anomaly appears in the western part of the 455 ACC, NW-oriented, 5.4-km long and 3-km wide with magnetic intensities between 84.5, and -2 456 nT. The map shows a contrast between the caldera (higher magnetic values) and the surrounding 457 areas (lower magnetic values). Different aeromagnetic domains or anomalies described in this 458 study are located either inside or outside of the ACC. 459 DISCUSSION 460 Tectonic implications 19 461 The regional tectonic setting has controlled the formation and evolution of the ACC. The 462 ACC formed on top of the Acoculco-El Rosario-Acoculco horst, which is bounded by the Apan- 463 Tlaloc and Tlaxco-Chignahuapan grabens (Fig. 67). Furthermore, some pre-existing tectonic 464 faults were used as pathways for dike intrusions and volcanism (aligned scoria cones parallel to 465 the NE-trending fault system), but also as part of the ring fault system that controlled the caldera 466 collapse and exerted the main control on the location of post-caldera vents (Fig. 2). The 467 southwestern caldera rim coincides with the NW-SE Manzanito fault, while the northern caldera 468 rim is marked by the Atotonilco scarp (Avellán et al., 2018; ) (Figs. 2 and 67). The sub-linear to 469 sub-circular collapsed structures form a 18 x 16 km rhombohedral-shape, steeply dipping towards 470 the central part of the ACC (Figs. 2 and 67). In fact, the left-lateral movement in the NE-SW 471 faults during the middle Miocene (García-Palomo et al., 2000; 2018) and/or the right-lateral 472 movement of the NW-SE faults could have been responsible for the fracturing and creation of the 473 space necessary to accommodate the magma reservoir beneath the caldera (pull-apart basins in 474 transtensional regimes; e.g., Bursik, 2009; Saxby et al., 2016). Activity of the NE-SW and NW- 475 SE regional systems continued after the caldera formation and modified the trace of the caldera 476 border, causing its displacement at several points (e.g., northeastern tip of the topographic caldera 477 rim, outcrops 27 and 28; Fig. 67B). Tectonic movements affected the interior of the caldera until 478 very recent time, causing displacements of intra-caldera blocks and disturbing the position of 479 intra-caldera volcanic vents and products. For example, E-W faulting in the central part of the 480 caldera (outcrops 24, 32 and 54; Fig. 67B) 481 Caldera reactivation during the emission of the 26 km3 Encimadas (~1.28 Ma) and 11 km3 482 of the Tecoloquillo ignimbrites (~0.6-0.76 Ma) was possibly originated near or along the E and 483 SW borders of the caldera, respectively (Figs. 2 and 3). Therefore, it would not be surprising if 484 the caldera used these faults systems to nucleate the rest of the ring fault that controlled the 20 485 collapse event, as it has occurred in other well documented calderas (see Aguirre-Díiaz et al., 486 2008; Martí et al., 2013; Molina et al., 2014). Despite the strong, selective erosion that has 487 affected parts of the area, the northern topographic caldera rim has not retreated significantly 488 from its original position. This may imply that caldera subsidence continued for a long time 489 (±500 ka, between Encimadas and Tecoloquillo eruptive events), dissecting younger rocks 490 emplaced close or through the ring faults. On the contrary, in the eastern side of the ACC, neither 491 the morphological nor the structural border are visible at surface, however, it is depicted in the 492 aeromagnetic map (Fig. 78). At the surface, the eastern border is not visible because it has been 493 coverburied by younger deposits and it is also likely an uneven collapse of the caldera. The E-W 494 intra-calderastructures fault system areis limited to the caldera interior and therefore isthey could 495 represent local deformation related to the resurgence of the caldera. 496 The NE-SW and NW-SE fault systems coexist in the ACC (Fig. 67B). This can be 497 explained by two different constructive phases. The first one took place in the Oligocene- 498 Miocene, ruled by ~NE-oriented extension, forming ~NW-SE dip-slip faults and ~NE-SW 499 sinistral strike-slip faults (left-lateral movement) ((Henry and Aranda-Gómez, 1992). The second 500 phase took place in the Pliocene-Pleistocene, and it is controlledruled by ~NW-oriented tectonic 501 extension, forming ~NW-SE dextral strike-slip faults (oblique and strike-slip faults with right- 502 lateral component) and ~NE-SW dip-slip faults (oblique and normal faults with a minor left- 503 lateral component; ) (García-Palomo et al., 2018). This is consistent with the left-stepping en 504 échelon geometry of the NW-SE Manzanito fault originated by right-lateral slip. Actually, this 505 latter phase of NW-oriented extension is still active (García-Palomo et al., 2018). Therefore, both 506 fault systems occur in the same region and are still active under the same stress regime. This is 507 possible because the NW-SE structures are acting as transfer faults of the NE-SW normal faults. 508 In order to demonstrate that both fault systems are moving under the same strain field, a synthetic 21 509 right dihedra diagram was created for the two pairs of conjugate fault sets, according to the mean 510 measured planes (Fig. 67B). This slip model implies an orthorhombic symmetry of faults, where 511 superposed areas divide the figure into the different compressional and tensional areas in the 512 ACC. 513 Aeromagnetic interpretation and geothermal implications 514 Previous studies used the aeromagnetic and gravimetric data of the Acoculco-Atotonilco 515 region obtained by Petroleos Mexicanos (PEMEX) in 1980 (García-Estrada, 2000; López- 516 Hernández et al., 2009). These authors interpreted a depression bounded by faultsstructure 517 coincident with gravimeetric data and low magnetic values with much higher values at the 518 interior defining the border of the Acoculco caldera. This coincides with our interpretation of the 519 airborne data of the Mexican Geological Survey obtained in 2000 displayed in the aeromagnetic 520 map of figure 78A. This map clearly shows an approximated match between the edge of the 18 x 521 16 km Acoculco caldera as reconstructed from the distribution of the Acoculco ignimbrite, the 522 co-ignimbritic breccia, and the Atotonilco scarp, and the Manzanito fault (Avellán et al, 2018; 523 this work) (Fig. 67). The aeromagnetic anomalies suggest a rough elliptical half-rounded limit of 524 the caldera that may be bounded to the east by the approximated venting location of the 525 Encimadas ignimbrite and to the southeast by the basaltic andesite la Paila scoria cone (Fig. 526 78A). The positive aeromagnetic anomaly located at the NE portion of the ACC is may be 527 produced by intrusive bodies underneath the Cretaceous limestones.the Zacatlán basaltic plateau 528 that Tthe Zacatlán basaltic plateau with high remanence and susceptibility. This basaltic plateau 529 is overlaid by the Encimadas ignimbrite and, both covering the Cretaceous limestones (Fig. 2). 530 The positive anomalies located to the SW and NW of the ACC may be are associated with 531 intrusive bodies linked to the ATVFthat gave place to monogenetic volcanoes of the ATVF 532 thandesitic units of Apan, the andesite-dacite units of Peñuela, and the late Miocene-late 22 533 Pleistocene basaltic andesitic units of ATVF (Figs. 2 and 78A). The distribution of these positive 534 anomalies are similar to the topographic features. The interior of the this tectonic depression the 535 caldera is not homogeneous but includes higher and lower magnetic values that might reveal the 536 presence of several small-scaleer horsts and grabens combined with the emplacement presence of 537 shallow intrusive bodies (i.e., laccoliths, sills) of basaltic and basaltic and intermediate andesitic 538 (i.e. sills) composition intruded inside the Cretaceous limestoneshosted inside the Cretaceous 539 limestones at depths (>1 km; (Fig. 78B).. These positive anomalies inside the caldera could have 540 fed the ca. 2.6 Ma Huistongo, 2.2 Ma Manzanito, 0.71 Ma Paila and <16 ka Cuatzitzingo 541 eruptions (Figs. 2, 3 and 7). at depth. rocks during the evolution of the caldera inside the 542 depression. The anomalies described in figure 78A have rough NE-SW orientations separating a 543 local horst on top of which the CFE exploratory wells were drilled, cutting skarns at ~790 m deep 544 (EAC-1), and argillitic limestones at 350 m deep (EAC-2) (Viggiano-Guerra et al., 2011). These 545 positive magnetic anomalies might represent intrusive bodies injected during the post-caldera 546 phase (Huistongo lavas 2.6 Ma), which may have an important role to keep active the geothermal 547 system active. Sosa-Ceballos et al. (2018) concluded that after the caldera collapse some 2.7 Ma 548 ago, the stress field on the ACC magma chamber and its immediate surroundings was modified. 549 Such change hindered magma migration through the collapsed reservoir and promoted lateral 550 tension zones where dikes and sills converge and eventually form new magma chambers. Due to 551 the heavily active tectonics deformed upper-crust in the Acoculco area, a new plumbing system 552 developed after the caldera collapse and gradually favored the ascent of deep seated, 553 independentphysically isolated peralkaline and calcalkaline mafic magmas. In addition, Sosa- 554 Ceballos et al. (2018) found that magma mixing-heating is the main magmatic process that 555 modified the ACC rock suite. Hence, the aeromagnetic anomalies might represent the horizontal 556 intrusion zones where mafic magma accumulates as a consequence of the upper crustal 23 557 deformation produced by the collapse (Fig. 2). These intrusion zones serve as heating elements of 558 magma reservoirs (that eventually might evolve into small magma chambers) and yield the 559 energy that sustain the Acoculco geothermal system (Fig. 2). When magma did not migrate 560 horizontally, it flowed upwards, sometimes ggetting ot tapped on its way to the surface evolving 561 to silicic compositions and in other cases got trapped as the aplitic dike dated at 183 ± 36 ka 562 found in EAC-1 (Sosa-Ceballos et al., 2018) (Table 1; Fig. 2). These young felsic intrusions have 563 reactivated the system providing heat for the hydrothermal activity and Holocene hydrothermal 564 explosions (Canet et al., 2015a and b). López-Hernández et al. (2009) also suggested the possible 565 presence of an intrusive body, whose top is at > 1000 m depth and that was not reached by EAC- 566 1 well. Although, we believe these intrusions also contribute to the geothermal system, their 567 volume and abundance are not yet known and probably can be resolved estimated by geophysical 568 methods to constraint their energy input. 569 A 3D aeromagnetic model of the area shows four anomalies (shallow intrusive bodies 570 likely mafic sills) intruded inside the Cretaceous limestones at depths >1 km (Fig. 7B). The CFE 571 exploratory wells were drilled between two of these anomalies where the wells cutting limestones 572 and skarns at different depths. These wells were drilled at a local structural high separated by two 573 depressions with NE-SW orientations (Fig. 6B). These depressions may be occupied by represent 574 intrusive bodies (laccoliths, sills) injected of basic and intermediate composition rocks 575 corresponding toduring the syn-caldera and early post-caldera rocks phase, in comparison with 576 the presence of more silicic rocks and sediments outside the caldera. The aplitic dike dated at 183 577 ± 36 ka supports the presence of young mafic intrusions that have reactivated the system 578 providing the heat for hydrothermal activity and Holocene hydrothermal explosions (Canet et al., 579 2015b). López-Hernández et al. (2009) also suggested the possible presence of an intrusive body, 24 580 whose top is at > 1000 m depth and that was not reached by EAC-1 well. The distribution of 581 magnetic anomalies inside the caldera also reveals the presence of tectonic lineaments with NE- 582 SW and NW-SE orientations, likely associated to horst and grabens blocks and coinciding with 583 those intra-caldera and caldera ring faults observed and measured in the field, suggesting that 584 such faults were used as vent zones for post-caldera volcanism (Figs. 2, 76 and 87 ). These faults 585 may have also favored the ascent of gas with isotopic compositions (N2/He, 3He/4He, 13C, 15N) of 586 both MORB- and arc-type signatures (Peiffer et al., 2014), and high values of CO2 and 3He/4He 587 ratio (R/Rair = 8.5), suggesting that suggest an active magmatic source at depth under the ACC 588 (Polak et al., 1982). 589 Volcanic evolution 590 Based in our synthetic map, updated refined volcanic stratigraphy, structural analysis, and 591 interpretation of the aeromagnetic anomalies interpretation is clear that ACC was emplaced 592 through the Cretaceous limestones, the Peñuela dacitic domes (~13 to 11 Ma), and the Zacatlán 593 basaltic plateau (Figs. 2 and 3). Rocks of Tthe ACC consists were divided intoof 30 units 594 corresponding formed into different stages of the caldera evolution from pre, syn, early-post, late- 595 post caldera and extra-caldera volcanism of the Apan-Tezontepec Volcanic Field (Avellán et al., 596 2018; Sosa-Ceballos et al., 2018). 597 Caldera formation (~2.7 Ma) 598 Prior to the formation of the caldera, the Puente and Terrerillos lava domes had been 599 extruded in the area (Fig. 89A). These domes of andesitic and dacitic composition and calc- 600 alkaline affinity were emplaced during the Pliocene between 3.9 and 3 Ma. After a few hundred 601 thousand years (0.3 Ma), a large amount of andesitic magma stagnated at depth preparing for a 602 major eruption. This calc-alkaline magma overpressured the encasing rock initiating the emission 603 of an andesitic pyroclastic density current (Acoculco ignimbrite). The continuous emission of the 25 604 ignimbrite eventually diminished the magmatic pressure and weakened the roof of the magma 605 chamber triggering its collapse (Fig. 89B). The collapse followed older structures as the NW-SE 606 Manzanito fault in the ignimbrite was dispersed from different sectors of the caldera structural 607 rim, along the Manzanito structure at the western and- southwestern parts and generated , the 608 Atotonilco structure at the northwest, and the new-formeda semi-circular caldera ring (Atotonilco 609 scarp) sector of the topographic caldera rim at the nNorth (Figs. 2 and 67B). The ignimbrite only 610 crops out in at the western- and southwestern and northern parts of the caldera interior because in 611 other locations it has beenwas covered by younger deposits. The collapse used pre-existing fault 612 systems that control the formed a 18 x 16 km asymmetric depression calderaof the ring fault 613 (Figs. 2 and 76). The used of older pre-existing faults (i.e., a long dike conduit) and fractures as 614 magma feeders facilitated a high discharge rate at the beginning of the eruption as discussed by 615 Costa and Martí (2016). This developed immediately into massive proportions, thus precluding 616 the formation of a vertical eruption column and the deposition of fallout deposits, which have not 617 been recognized neither outside the caldera nor in the intra-caldera wells. The occurrence of a co- 618 ignimbrite breccia found in relative small, isolated areas at the north, west and southwestern parts 619 of the caldera rim points to emissions sites of the ignimbrite (Fig. 67B). In fact, the ring fault 620 scarps and the topographic caldera rim are clearly visible at the western half of the caldera, but 621 they are hidden at the eastern side, suggest that caldera-collapse could have been more intense in 622 the western sector than in the eastern sector. We cannot inform on how the caldera-forming 623 eruption progressed, as most if its corresponding deposits have been eroded outside the caldera or 624 are now totally hidden by post-caldera rocks at the interior of the caldera depression. However, 625 the presence of a large number of normal faults with different orientations affecting intra-caldera 626 rocks suggests that caldera-collapse could have occurred in a piecemeal-trapdoor fashion through 627 several stages of volcanism. 26 628 Early-post caldera phase (2.6-2.2 Ma) 629 After the caldera formation followed a relatively short (~0.1 Ma) quiescence period 630 during which an intra-caldera shallow lake developed with lacustrine sedimentation (Fig. 56B). 631 The early post-caldera volcanism is represented by eruptions that occurred mainly in the interior 632 of the caldera with a minimum total volume of 27 km3 (Fig. 89C). This volcanism vented along 633 the ring faults and intra-caldera normal faults that formed or acted during caldera collapse to then 634 be reactivated by the extensional tectonics that have affected the area until present (García- 635 Palomo et al., 2018). These modifications of the local stress field allowed the ascent of 636 peralkaline basaltic and basaltic trachyandesite magmas that were generated by partial melting of 637 a metaszomatized mantle, genetically unrelated to the calc-alkaline magmas that dominated the 638 pre-caldera and caldera activities (Sosa-Ceballos et al., 2018). Both suites of magma mixed and 639 formed the early post-collapse dome complexes on the ring of the caldera. 640 Late-post caldera phase (~2.0 - ~0.016 Ma) 641 After a quiescent period of around 0.2 Ma the caldera entered a new phase of volcanism 642 along the caldera rim dominated by the emission of domes and lava flows, two ignimbrites, and 643 four scoria cones (Fig. 89D). An extended period of intermittent volcanism occurred between 2 644 and 1.3 Ma with the occurrence of at least 13 effusive eruptions along the caldera rim. Between 2 645 and 1.6 Ma eruptions emitted bimodal andesitic (Coal, Tal, Cual units) and rhyolitic (Alrd, Lrd 646 units) products to then erupted rhyolitic effusive magmas until 1.28 Ma (Ard unit) at the 647 southwestern part of the caldera rim. At about the same time (1.27 Ma) a rhyolitic magma 648 explosively flared explosion occurred at the eastern part of the caldera dispersing to the through 649 the eastern part of the caldera dispersing to the east and northeast the Encimadas ignimbrite (26 650 km3) Encimadas ignimbrite (26 km3). Encimadas was apparently erupted from a previous sealed 651 portion of the caldera (tilted blocked to the west of the piece meal collapse) since no Acoculco 27 652 ignimbrite has been found eastern of the caldera. We assume that magmatic overpressure in a 653 shallow magma chamber triggered the rhyolitic eruption that represents the second largest event 654 of the caldera. Immediately after this major eruption occurred four rhyolitic eruptions in the S- 655 SW parts of the caldera between 1.27 and 1.06 Ma. These eruptions particularly emitted the 656 voluminous Ailitla coulée dome to the south (Arcd; 12.4 km3), the Cabezas lava to the southwest 657 (Crl), a minor explosive event that dispersed small-volume pyroclastic density currentsPDCs 658 (Msf), and the Minilla coulée come to the west (Mrcd). These effusive eruptions precluded 659 preceded the occurrence of another explosive rhyolitic event that occurred some 0.2 Ma later. 660 This eruption vented at the southwestern edge of the caldera rim (0.6-0.76 Ma) at the intersection 661 with the NW-SE Manzanito fault. This eruption dispersed the Tecoloquillo rhyolitic ignimbrite 662 (11 km3) to the southwest of the caldera rim ending with the extrusion of a rhyolitic summit 663 dome. The Tecoloquillo ignimbrite represents the third largest explosive event of the ACC after 664 which activity associated to the caldera apparently declined for a period of circa 0.5 Ma. 665 However, renewed activity of the caldera started 0.071 Ma with the emplacement of the La Paila 666 (Plc) at the southeast, the Tuliman scoria cone and lava flow (0.063 Ma) at the northwest, and the 667 Cuatzitzinguito (16 ka) and Cuatzitzingo ca. 0.01 Ma (16 ka) scoria cones at the southeast. These 668 magmas have erupted basaltic andesite to basaltic trachyandesite products (Sosa-Ceballos et al., 669 2018). 670 During this late post-caldera period the peralkaline magma suite gradually dominated 671 over the calc-alkaline suite (Sosa-Ceballos et al., 2018) (Figs. 89D-C). Thise mixing process 672 resulted in an undistinguishable unique source for the early post-collapse magmas, whereas trace 673 elements geochemistry suggests a relatively homogeneous source for the late post-collapse 674 rhyolites (Sosa-Ceballos et al., 2018). The total volume of magma erupted during late post- 675 caldera phase was ~90 km3. We believe that during this phase of the caldera evolution began the 28 676 resurgence of the caldera floor as supported suggested by uplifted and tilted lacustrine sediments 677 (250°/30SE) exposed inside the caldera (Fig. 2) covering the Acoculco ignimbrite. The 678 morphology of the caldera floor suggests that the resurgence was more important in the 679 southwestern portion, which is in accordance with shallow intrusions of magma revealed by 680 aeromagnetic data that could cause the caldera resurgence. According to Sosa-Ceballos et al. 681 (2018) these shallow magma intrusions were originated by post-caldera deformation that 682 promoted the formation ofa swarm of dikes and sills and dykes above the collapsed reservoir and 683 might be the heat source of the Acoculco geothermal system. 684 Finally, the area covered by the AAC products is ca. 856 km2 with a total minimum 685 volume ca. 143 km3. By assuming that the present extension of deposits is minimum, we 686 calculated that activity was dominated by nearly 80 % of effusive volcanism with three major 687 bursts of explosive volcanism that represent 20% of the ACC. As summarized above Acoculco 688 has been a site of persistent volcanism since 2.7 Ma that include the presence of a young 689 intrusion (183 ± 36 ka) at the bottom of the EAC-1 well. This intense magmatism, as well as 690 Holocene hydrothermal explosions (Canet et al., 2015a; 2015b), and geothermal manifestations 691 (Peiffer et al., 2014) indicate that the complex is still active and could represent a site to develop 692 a geothermal energyfield. 693 CONCLUSIONS 694 The Acoculco collapse caldera was originated at about ~2.7 Ma ago in response to the 695 eruption of 127 km3 of the andesitic ignimbrite of the Acoculco unit. The caldera collapse episode 696 occurred in part along a pre-existing regional NE-SW trending faults (southwestern, western and 697 northwestern sectors), but also along newly formed ring faults (northern sector and probably the 698 eastern sector, but this is not visible). The collapse occurred in a piecemeal-trapdoor fashion, in 699 which intra-caldera blocks bounded by pre-existing or newly formed normal faults collapsed 29 700 gravitationally collapsed in a partly emptied to the associated magma chamber. After the 701 formation of the caldera, volcanic activity stopped for a while permitting the formation of an 702 intra-caldera lake as suggested by, which sediments covering part ofed the caldera-forming 703 ignimbrite. The rRenewedal of the volcanic activity produced the emplacedment of several 704 domes and lava flowsvolcanic units along the caldera borders and intra-caldera faults that entirely 705 filledmodified the caldera depression. Post-caldera volcanism also affected the areas around the 706 caldera and was controlled by the main regional fault systems that also influenced the formation 707 of the caldera. The ACC was formed at the intersection of NE-SW and NW-SE fault systems. 708 These fault systems have controlled the formation and evolution of the ACC, the regional 709 subsidence, and venting of syn- and post-caldera volcanism inside and outside the caldera. 710 The presence of positive aeromagnetic anomalies associated to intrusive bodies (sills and 711 dykes) thermal anomaly at the interior of the Acoculco caldera makes it an interesting target for 712 geothermal exploration. This geothermal anomaly is probably related to the existence of more 713 than ~2.7 Ma of post-caldera volcanism and the magma trapped in horizontal intrusions, as 714 suggested by aeromagnetic data. 715 ACKNOWLEDGEMENTS 716 This study was partially funded by the Centro Mexicano de Innovación en Energía 717 Geotérmica (CeMIE-Geo) project P15 and GeMex 4.4. to J.L. Macías. We thank F. Mendiola, G. 718 Reyes-Agustín and S. Cardona for their technical support during the laboratory analyses and D.E. 719 Torres-Gaytan for his support during the aeromagnetic map generation. We appreciate the 720 discussions and input provided by C. Arango, C. Canet and J. Marti is grateful for the MECD 721 (PRX16/00056) grant. We appreciated the constructive comments by J. Aranda and two 722 anonymous reviewers. 723 REFERENCES CITED 30 724 Afanasyev, A., Costa, A., Chiodini, G., 2015. Investigation of hydrothermal activity at Campi 725 Flegrei caldera using 3D numerical simulations: Extension to high temperature 726 processes. Journal of Volcanology and Geothermal Research, 299, 68-77. 727 Aguirre-Dı́az, G.J., López-Martı́nez, M., 2001. The Amazcala caldera, Queretaro, Mexico. 728 Geology and geochronology. Journal of Volcanology and Geothermal Research, 111(1-4), 729 203-218. 730 Aguirre‐Díaz, G.J., Nieto‐Obregón, J., Zúñiga, F.R., 2005. Seismogenic basin and range and 731 intra‐arc normal faulting in the central Mexican Volcanic Belt, Querétaro, México. Journal of 732 Geology, 40(2), 215-243. 733 Aguirre-Díaz, G.J., Labarthe-Hernández, G., Tristán-González, M., Nieto-Obregón, J., Gutiérrez- 734 Palomares, I., 2008. The ignimbrite flare-up and graben caldera of the Sierra Madre 735 Occidental, México. In: Gottsmann J, Martí J. (Eds.), Caldera volcanism: analysis, modeling 736 and response. Elsevier, Amsterdam, 143–174. 737 Alatriste-Vilchis, D., Campos-Enriquez, O., Huizar-Alvarez, R., Marines-Campos, R., 2005. La 738 estructura sub-superficial de la subcuenca de Tecocomulco. La Laguna de Tecocomulco Geo- 739 Ecologia de un desastre. Universidad Nacional Autónoma de México, Instituto de Geología, 740 Publicación especial, 3, 33-48. 741 Angelier, J., 1990. Inversion of field data in fault tectonics to obtain the regional stress—III. A 742 new rapid direct inversion method by analytical means. Geophysical Journal International, 103 743 (2), 363-376. 744 Aranda-Gómez, J.J., McDowell, F.W., 1998. Paleogene extension in the southern Basin and 745 Range province of México: Syndepositional tilting of Eocene red beds and Oligocene volcanic 746 rocks in the Guanajuato Mining District. International Geology Review, 40 (2), 116-134. 747 Avellán, D.R., Macías, J.L., Layer, P.W., Cisneros, G, Sánchez-Núñez, J.M., Gómez- 748 Vasconcelos, M.G., Pola, A., Sosa-Ceballos, G., García-Tenorio, F., Reyes-Agustín, G., 749 Osorio-Ocampo, S., García-Sánchez, L., Mendiola, I.F., Marti, J., López-Loera H., Benowitz, 750 J., 2018. Geology of the late Pliocene – Pleistocene Acoculco caldera complex, eastern Trans- 751 Mexican Volcanic Belt (México). Journal of Maps, 15(2), 8-18. 752 Bibby, H.M., Caldwell, T.G., Davey, F.J., Webb, T.H., 1995. Geophysical evidence on the 753 structure of the Taupo Volcanic Zone and its hydrothermal circulation. Journal of 754 Volcanology and Geothermal Research, 68 (1-3), 29-58. 31 755 756 Bursik, M., 2009. A general model for tectonic control of magmatism: Examples from Long Valley Caldera (USA) and El Chichón (México). Geofísica internacional, 48(1), 171-183. 757 Caballero, M., Lozano, S., Ortega, B., Urrutia, J., Macias, J.L., 1999. Environmental 758 characteristics of Lake Tecocomulco, northern basin of México, for the last 50,000 759 years. Journal of Paleolimnology, 22 (4), 399-411. 760 Calcagno, P., Evanno, G., Trumpy, E., Gutiérrez-Negrín, L.C., Macías, J.L., Carrasco-Núñez, G., 761 Liotta, D., the GEMex T3.1 team, 2018. Preliminary 3-D geological models of Los Humeros 762 and Acoculco geothermal fields (Mexico) – H2020 GEMex Project. Advances in Geosciences, 763 1, 1-13. 764 Campos-Enríquez, J.O., Alatriste-Vilchis, D.R., Huizar-Álvarez, R., Marines-Campos, R., 765 Alatorre-Zamora, M.A., 2003. Subsurface structure of the Tecocomulco sub-basin 766 (northeastern México basin), and its relationship to regional tectonics. Geofísica Internacional, 767 42 (1), 3-24. 768 Canet, C., Hernández-Cruz, B., Jiménez-Franco, A., Pi, T., Peláez, B., Villanueva-Estrada, R.E., 769 Salinas, S., 2015a. Combining ammonium mapping and short-wave infrared (SWIR) 770 reflectance spectroscopy to constrain a model of hydrothermal alteration for the Acoculco 771 geothermal zone, Eastern México. Geothermics, 53, 154-165. 772 Canet, C., Trillaud, F., Prol-Ledesma, R.M., González-Hernández, G., Peláez, B., Hernández- 773 Cruz, B., Sánchez-Córdova, M.M., 2015b. Thermal history of the Acoculco geothermal 774 system, eastern México: Insights from numerical modeling and radiocarbon dating. Journal of 775 Volcanology and Geothermal Research, 305, 56-62. 776 777 778 779 Castro-García, A., Córdoba, D., 1994. Estratigrafía del área volcánica de Tulancingo, Hidalgo, México. Revista de Investigación, Universidad Autónoma del Estado de Hidalgo, 1, 26-39. Costa, A., Martí, J., 2016. Stress field control during large caldera-forming eruptions. Frontiers in Earth Science, 4 (92), 1-13. 780 De Cserna, Z.D., De la Fuente-Duch, M., Palacios-Nieto, M., Triay, L., Mitre-Salazar, L. M., 781 Mota-Palomino, R., 1988. Estructura geológica, gravimetría, sismicidad y relaciones 782 neotectónicas regionales de la cuenca de México. Boletín del Instituto de Geología, 783 Universidad Nacional Autónoma de México, 104, 71. 784 785 De la Cruz-Martínez, V., Castillo-Hernández, D., 1986. Geología de la zona geotérmica de la caldera de Acoculco, Puebla. Geotermia, Revista Mexicana de Geoenergía, 2 (3), 245-254. 32 786 Di Napoli, R., Martorana, R., Orsi, G., Aiuppa, A., Camarda, M., De Gregorio, S., Candela E.G., 787 Luzio D., Messina, N., Pecoraino, G., Bitetto, M., de Vita, S., Valenza M., 2011. The structure 788 of a hydrothermal system from an integrated geochemical, geophysical, and geological 789 approach: The Ischia Island case study. Geochemistry, Geophysics, Geosystems, 12 (7), 1-25. 790 De Vicente, G., Muñoz, A., Giner, J.L., 1992. El uso del método de los diedros rectos: 791 implicaciones a partir del modelo de deslizamiento del análisis poblacional de fallas. Revista 792 Sociedad Geológica de España, 5, 3-4. 793 Ellis, R.G., De Wet, B., MacLeod, I.N., 2012. Inversion of magmatic data for remanent and 794 induced sources: Auxiliary Information in Geophysical Inversion. Presented at the ASEG 795 Inversion Forum, extended abstract. 796 797 García-Estrada, G.H., 2000. Gravimetría y magnetometría regional del proyecto Acoculco, Puebla. CFE-GPG internal report DEX-ACO19/2000, 57. 798 García-Palomo, A., Macías, J.L., Garduño, V.H., 2000. Miocene to Recent structural evolution 799 of the Nevado de Toluca volcano region, central México. Tectonophysics, 318 (1), 281-302. 800 García-Palomo, A., Macías, J.L., Tolson, G., Valdez, G., Mora, J.C., 2002. Volcanic stratigraphy 801 and geological evolution of the Apan región, east-central sector of the Trans-Mexica Volcanic 802 Belt. Geofísica Internacional, 41 (2), 133-150. 803 García-Palomo, A., Macías, J.L., Jiménez, A., Tolson, G., Mena, M., Sánchez-Núñez, J.M., 804 Arce, J.L, Layer, P.W., Santoyo, M.A., Lermo-Samaniego, J., 2018. NW-SE Pliocene- 805 Quaternary extension in the Apan-Acoculco region, eastern Trans-Mexican Volcanic Belt. 806 Journal of Volcanology and Geothermal Research, 349, 240-255. 807 García-Sánchez, L., Macías J.L., Arce, J.L., Sosa-Ceballos, G., Garduño-Monroy, V.H., 808 Saucedo, R., Avellán, D.R., Rangel, E., Layer, P.W. Lopez-Loera, H., Rocha, V.S., Cisneros, 809 G., Reyes-Agustín; G., Jiménez, A., 2016. Genesis and evolution of the Cerro Prieto Volcanic 810 Complex, Baja California, Mexico. Bulletin of Volcanology, 79 (44), 1-28. 811 García-Tovar, G.P., Martínez-Serrano, R.G., Solé, J., Correa-Tello, J.C., Nuñez-Castillo, E.Y., 812 Guilluo, H., Monroy-Rodríguez, E., 2015. Geología, geocronología y geoquímica del 813 vulcanismo Plio-Cuaternario del Campo Volcánico Apan-Tecocomulco, Faja Volcánica 814 Trans-Mexicana. Revista Mexicana de Ciencias Geológicas, 32 (1), 100-122. 815 Geyer, A., Martí, J., 2014. A short review of our current understanding of the development of 816 ring faults during collapse caldera formation. Frontiers in Earth Science, Volcanology, 2, 1-22. 33 817 818 Henry, C.D., Aranda-Gomez, J.J., 1992. The real southern Basin and Range: Mid-to late Cenozoic extension in México. Geology, 20 (8), 701-704. 819 Hernández-Javier, I., 2008. Geología y geomorfología volcánica de la región de los yacimientos 820 de obsidiana en Otumba en el sector norte de la Sierra Nevada de México. Bacherlor thesis, 821 Universidad Nacional Autónoma de México. 822 Hiriart, G., Gutiérrez, N. L., Quijano, L.L., Ornelas, C.A., Espíndola, S., Hernández, I., 2011. 823 Evaluación de la energía geotérmica en México. Informe para el Banco Interamericano de 824 Desarrollo y la Comisión Reguladora de Energía, México. 825 Huízar-Álvarez, R., Campos-Enríquez, J.O., Lermo-Samaniego, J., Delgado-Rodríguez, O., 826 Huidobro-González, A., 1997. Geophysical and hydrogeological characterization of the sub- 827 basins of Apan and Tochac (México Basin). Geofísica Internacional, 36, 217-234. 828 Izquierdo, G., Aragón, A., Portugal, E., Arellano, V.M., De León, J., Álvarez, J., 2006. 829 Mineralogía de la zona mineralizada de sílice-epidota (ZMSE) del yacimiento geotérmico de 830 Cerro Prieto, B. C., México. Geotermia, Revista Mexicana de Geoenergia, 19, 2-12. 831 832 Lorenzo-Pulido, C., Armenta-Flores, M., Ramírez-Silva, G., 2010. Characterization of the 833 Acoculco Geothermal Zone as a HDR System. Geothermal Resources Council Transactions, 834 34, 369–372. 835 836 837 838 Johnson, C.A., Harrison, C.G.A., 1990. Neotectonics in central México. Physics of the Earth and Planetary Interiors, 64, 187-210. Layer, P.W., 2000. Argon-40/argon-39 age of the El’gygytgyn impactevent, Chukotka, Russia. Meteorite Planetary Science, 35, 591–599. 839 Layer, P.W., García-Palomo, A., Jones, D., Macías, J.L., Arce, J.L., Mora, J.C., 2009. El 840 Chichón volcanic complex, Chiapas, México: stages of evolution based on field mapping and 841 40Ar/39Ar geochronology. Geofisica Internacional, 48, 33–54. 842 Ledezma-Guerrero, O., 1987. Hoja Calpulalpan 14Q-h (3), con Resumen de la geología de la 843 hoja Calpulalpan, Estados de México, Puebla, Tlaxcala e Hidalgo. Quetzalapa region, southern 844 México. Revista Mexicana de Ciencias Geológicas, 15, 167-180. 845 Lermo, J., Antayhua, Y., Bernal, I., Venegas, S., Arredondo, J., 2009. Monitoreo sísmico en la 846 zona geotérmica de Acoculco, Puebla, México. Geotermia, Revista Mexicana de Geoenergía, 847 22 (1), 40-58. 34 848 849 Lipman, P.W., 2000. Calderas. In: Sigurdsson, H., Houghton, B.F., McNutt, S.R., Rymer, H., Stix, J. (Eds.), Encyclopedia of Volcanoes. Academic press, San Diego, 331-343. 850 López-Hernández, A., 2009. Evolución volcánica del complejo Tulancingo-Acoculco y su 851 sistema hidrotermal, Estados de hidalgo y Puebla, México. Ph.D. thesis, Universidad Nacional 852 Autónoma de México. 853 Lopez-Hernandez, A., Castillo-Hernandez, D., 1997. Exploratory drilling at Acoculco, Puebla, 854 México: a hydrothermal system with only nonthermal manifestations (No. CONF-971048). 855 Geothermal Resources Council, Davis, CA (United States). 856 López-Hernández, A., García-Estrada, G., Aguirre-Díaz, G., González-Partida, E., Palma- 857 Guzmán, H., Quijano-León, J., 2009. Hydrothermal activity in the Tulancingo-Acoculco 858 Caldera Complex, central México - Exploratory studies. Geothermics, 38, 279-293. 859 López-Hernández, A., Martínez, E.I., 1996. Evaluación volcanológica y estructural de la zona 860 geotérmica de Acoculco, Puebla, y su relación con la anomalía termal detectada en el pozo 861 EAC-1. CFE-GPG Internal Report OGL-AC-11/96. 862 Macías, J.L., Arce, J.L., García-Tenorio, F., Layer, P.W., Rueda, H., Reyes-Agustin, G., 863 Avellán, D., 2012. Geology and geochronology of Tlaloc, Telapón, Iztaccíhuatl, and 864 Popocatépetl volcanoes, Sierra Nevada, central México. Field Guides, 25, 163-193. 865 866 867 868 869 870 Martí, J., Geyer, A., Folch, A., Gottsmann, J., 2008. A review on collapse caldera modelling. Developments in Volcanology, 10, 233-283. Martí, J., Geyer, A., Aguirre-Diaz, G., 2013. Origin and evolution of the Deception island caldera (South Shetland Islands, Antarctica). Bulletin of Volcanology, 75 (6), 732. McDougall I., Harrison, T.M., 1999. Geochronology and thermochronology by the 40Ar/39Ar method, 2nd edition. Oxford University Press, New York. 871 Molina, F., Martí, J., Aguirre, G., Vega, E., Chavarría, L., 2014. Stratigraphy and structure of the 872 Cañas Dulces caldera (Costa Rica). Geological Society of America Bulletin, 126 (11-12), 873 1465-1480. 874 875 Mooser, F., Ramírez, M.T., 1987. Faja Volcánica Transmexicana: Morfoestructura, tectónica y vulcanotectónica. Boletín de la Sociedad Geológica Mexicana, 48 (2), 69-73. 876 Pardo, M., Suárez, G., 1995. Shape of the subducted Rivera and Cocos plates in southern 877 México: Seismic and tectonic implications. Journal Geophysical Research, 100 (12), 357-12. 35 878 Peiffer, L., Bernard-Romero, R., Mazot, A., Taran, Y.A., Guevara, M., Santoyo, E., 2014. Fluid 879 geochemistry and soil gas fluxes (CO 2–CH 4–H 2 S) at a promissory Hot Dry Rock 880 Geothermal System: The Acoculco caldera, México. Journal of Volcanology and Geothermal 881 Research, 284, 122-137. 882 Pérez‐Campos, X., Kim, Y., Husker, A., Davis, P.M., Clayton, R.W., Iglesias, A., Pacheco, J., 883 Singh, S., Constantin, M.V., & Gurnis, M., 2008. Horizontal subduction and truncation of the 884 Cocos Plate beneath central Mexico. Geophysical Research Letters, 35(18). 885 Polak, B.G., Prasolov, E.M., Kononov, V.I., Verkhocskiy, A.B., González, A., Templos, L. A., 886 Mañón, A., 1982. Isotopic composition and concentration of inert gases in Mexican 887 hydrothermal systems (genetic and applied aspects). Geofísica Internacional, 21 (3), 193-227. 888 Rocha, S., Jiménez, E., Palma, H., 2006. Propuesta para dos pozos exploratorios en el proyecto 889 geotérmico de Acoculco, Puebla. Comisión Federal de Electricidad, internal report OGL- 890 ACO-03/06, unpublished. 891 Renne P.R., Mundil, R., Balco, G., Min, K., Ludwig, K.R., 2010. Joint determination of 40K 892 decay constants and 40Ar*/40K for the Fish Canyon sanidine standard, and improved 893 accuracy for 40Ar/39Ar geochronology. Geochimica et Cosmochimica Acta, 74, 5349–5367 894 Rueda, H., Macías, J.L., Arce, J.L., Gardner, J.E., Layer, P.W., 2013. The~ 31ka rhyolitic 895 Plinian to sub-Plinian eruption of Tlaloc Volcano, Sierra Nevada, central México. Journal of 896 Volcanology and Geothermal Research, 252, 73-91. 897 Ryan, W.B., Carbotte, S.M., Coplan, J.O., O'Hara, S., Melkonian, A., Arko, R., Weissel, R.A., 898 Ferrini, V., Goodwillie, A., Nitsche, F., Bonczkowski, J., 2009. Global multi‐resolution 899 topography synthesis. Geochemistry, Geophysics, Geosystems, 10 (3). 900 901 Saxby, J., Gottsmann, J., Cashman, K., & Gutiérrez, E., 2016. Magma storage in a strike-slip caldera. Nature communications, 7, 12295. 902 Siebe, C., Macias, J.L., Abrams, M., Rodriguez, S., Castro, R., Delgado, H., 1995. Quaternary 903 explosive volcanism and pyroclastic deposits in east central México: implications for future 904 hazards. In: Chacko, J., Whitney A. (Eds.), Guidebook of geological excursions, in 905 conjunction with the Annual Meeting of the Geological Society of America. New Orleans, 906 Louisiana, 1-48. 907 Sosa-Ceballos, G., Macías, J.L., Avellán, D.R., Salazar-Hermenegildo, N., Boijseauneau-López, 908 M.E., Pérez-Orozco, J.D., 2018. The Acoculco Caldera Complex magmas: Genesis, evolution 36 909 and relation with the Acoculco geothermal system. Journal of Volcanology and Geothermal 910 Research, 358, 288-306. 911 Steiger, R., Jäger, E., 1977. Subcommission on geochronology: convention on the use of decay 912 constants in geo-and cosmochronology. Earth and planetary science letters, 36 (3), 359-362. 913 Suter, M., 1984. Cordilleran deformation along the eastern edge of the Valles–San Luis Potosí 914 carbonate platform, Sierra Madre Oriental fold-thrust belt, east-central Mexico. Geological 915 Society of America Bulletin, 95(12), 1387-1397.Suter, M., 1991. State of stress and active 916 deformation in México and western Central America. Neotectonics of North America, 1, 401- 917 421. 918 Suter, M., Carrillo-Martínez, M., Quintero-Legorreta, O., 1996. Macroseismic study of shallow 919 earthquakes in the central and eastern parts of the trans-Mexican volcanic belt, 920 México. Bulletin Seismological Society of America, 86 (6), 1952-1963. 921 922 923 924 Tello-Hinojosa, E., 1986. Geoquímica de la zona geotérmica de la caldera de Acoculco, Puebla. CFE-GPG internal report 34/86, 15. Urrutia-Fucugauchi, J., Flores-Ruiz, J., 1996. Bouger gravity anomalies and regional crustal structure in central Mexico. International Geology Review, 38, 176–194. 925 Viggiano-Guerra, J.C., Flores-Armenta, M., Ramírez-Silva, G.R., 2011. Evolución del sistema 926 geotérmico de Acoculco, Pue., México: un estudio con base en estudios petrográficos del pozo 927 EAC-2 y en otras consideraciones. Geotermia, Revista Mexicana de Geoenergia, 24 (1), 14- 928 24. 929 Zúñiga, F.R., Pacheco, J.F., Guzmán-Speziale, M., Aguirre-Dıaz, G.J., Espındola, V.H., Nava, 930 E., 2003. The Sanfandila earthquake sequence of 1998, Querétaro, México: activation of an 931 undocumented 932 Belt. Tectonophysics, 361 (3), 229-238. fault in the northern edge of central Trans-Mexican Volcanic 933 934 FIGURE CAPTIONS 935 Figure 1. A) Regional tectonic configuration of Central México showing the active subduction of 936 the Rivera and Cocos Plates beneath the North American Plate at the Middle-American Trench. 937 The dotted black line depicts the boundary of the Trans-Mexican Volcanic Belt (TMVB) and the 938 location of the Acoculco Caldera Complex and other volcanoes (LP, La Primavera; C, Fuego de 37 939 Colima; T, Tancítaro; Am, Amealco; Hc, Huichapan; To, Nevado de Toluca; P, Popocatepetl; M, 940 Malinche; O, Pico de Orizaba; and Hm, Humeros). The shaded relief map and bathymetric data 941 was acquired and modified from Ryan et al. (2009). 942 Figure 2. Simplified geological map of the Acoculco Caldera Complex after Avellán et al. 943 (2018), Sosa-Ceballos et al. (2018), and Calcagno et al. (2018). The map contains the distribution 944 of 30 stratigraphic units and a composite stratigraphic column supported by 24 new 945 ages and two 14C dates (Fig. 3). This map shows towns, main roads, the location of stratigraphic 946 logs, samples with age determinations, faults, the zone of intense hydrothermal alteration, and 947 the two CFE exploratory wells. The DEM was obtained from the INEGI topographic data with a 948 horizontal resolution of 15-m2. The geologic profile at the bottom correlates the surface units of 949 this map with the CFE stratigraphic logs of the EAC-1 and EAC-2 wells with the location of 950 shallow intrusions. 951 Figure 3. Detailed stratigraphic column of the units and eruptive stages of the Acoculco Caldera 952 Complex composed of the 30 units displayed marked in the geological map of figure 2. To the 953 right appear the plateau age of units, the plateau age sorted by age, and their estimated total 954 volume. 955 Figure 4. Representative 956 the Acoculco Caldera Complex (see Table 1). Plateau and Isochron ages, where calculated, are 957 reported at the 1-sigma level. For Ac90, the 6 single-step fusion results are shows as a single 958 probability density histogram 959 rocks of the Acoculco Caldera Complex (see Table 1). 960 Figure 5. Photos of selected outcrops of the Acoculco andesitic ignimbrite (Aai) with its different 961 lithological facies. A, Aspect of the welded massive structure of the ignimbrite with pumice 962 fragments embedded in a light-yellowish ash matrix. Hand lens for scale. B, Lacustrine deposits 40Ar/39Ar 40Ar/39Ar age spectra and inverse isochron diagrams of dated rocks of 40Ar/39Ar age spectrums and inverse isochron diagrams of dated 38 963 that partly covereding the Aai with parallel plane stratification. They are made of interbedded 964 massive fine sandy beds with laminar-stratified clayed beds. Notice that the succession is gently 965 tilted. C, Co-ignimbrite lag breccia (mlBr) of the Acoculco ignimbrite transected by E-W 966 structures and NW-SE faults. Notice that the lava lithics are supported by a light-brown to 967 yellowish ashy matrix with pumice. D, Co-ignimbrite lag breccia (mlBr) at the bottom of the 968 outcrop in contact with the Acoculco ignimbrite. The color change marks a difference in the 969 texture and lithological components of the deposits. 970 Figure 6. 971 Caldera Complex (see Table 1). Plateau and Isochron ages, where calculated, are reported at the 972 1-sigma level. 973 Figure 7. A, Regional structural map modified from García-Palomo et al. (2018). Dark gray areas 974 correspond to horsts and light gray areas correspond to grabens. Dike, scoria cones and 975 slickensides symbols show extension direction. B, Structural map of the Acoculco Caldera 976 Complex showing the main structural features, field sites (outcrops) and the caldera rim (this 977 study). C, The fault planes were plotted on stereographic projections in the lower hemisphere. 978 The trend-frequency polar diagramstereogram (rose diagram) was obtained with Rozeta 2.0, and 979 the fault planes and poles were plotted with Stereonet 9. D, A slip model showing an 980 orthorhombic symmetry of NW- and NE- oriented faults moving under the same strain field. 981 Black zones correspond to 100% compatible compression, and white zones 100% tension areas. 982 Figure 8. A, Image of the Magnetic Field Reduced to the Pole (MFRP) overlaid on a DEM of the 983 ACC region. The MFRP values are represented by the horizontal bar color at the base of the 984 figure with red colors standing for high values and the blue ones for low values of the magnetic 985 intensities. The black dots represent the CFE wells. B, A 3D inversion model of the magnetic 986 susceptibility produced with the Geosoft´s VOXI software by using the magnetic vector inversion 40Ar/39Ar age spectra and inverse isochron diagrams of dated rocks of the Acoculco 39 987 method (Ellis et al., 2012). In this image four anomalies related to intrusive bodies occur at 988 depths between 1,000 and >2,500 m below the surface. 989 Figure 9. Eruptive stages during the formation of the Acoculco Caldera Complex described as 990 pre-caldera (A), syn-caldera (B), early post-caldera (C) late post-caldera (D), and still late post- 991 caldera and extra-caldera (E). After the caldera collapse two episodes of caldera reactivation 992 generated the Encimadas (Eri), and Tecoloquillo (Tr1) rhyolitic ignimbrites. Outside of the 993 caldera rim, Tthe products of the Acoculco caldera are interbbededinterbedded with deposits of 994 the Apan-Tezontepec volcanic field (extra caldera volcanism). 995 LIST OF TABLES 996 Table 1. List of dated samples from rock units of the Acoculco Caldera Complex. See text for 997 map units. All ages (in ka) are reported at the 1-sigma level. For most samples, the 40Ar/39Ar 998 plateau age is the interpreted age (shown in bold) of the unit. Sample Aco901 showed evidence 999 of excess argon, so the isochron age is the interpreted age. The age of sample Ac90 is a weighted 1000 mean age of 6 single-step fusions, and for sample Ac94a, no plateau or isochron could be 1001 calculated. MSWD: Mean Square Weighted Deviation. See text for analytical methods and 1002 standards used. List of rock units of the Acoculco Caldera Complex dated in this study and their 1003 40Ar/39Ar 1004 were calculated using the constants of Steiger and Jäeger (1977). Analyses performed at the 1005 Geochronology Laboratory of the University of Alaska at Fairbanks, USA. 1006 Table 2. Radiocarbon date of paleosol sample performed during study. The location is in UTM 1007 coordinates. Analyses performed at the International Chemical Analysis Inc. Florida, USA. ages. , pPreferred results are listed in the plateau age (ka) column in bold. All ages are 1008 40 Highlights  Acoculco caldera complex lies at the intersection of NE-SW and NW-SE fault systems  The complex has emplaced ~143 km3 of magma since 2.7 Ma.  Acoculco has erupted andesitic to rhyolitic magmas from 2.7 to ca. 0.016 Ma  Young intrusions are likely providing the potential geothermal source 1 Eruptive chronology of the Acoculco caldera complex – a resurgent 2 caldera in the eastern Trans-Mexican Volcanic Belt (México) 3 by 4 5 6 7 Denis Ramón Avellán1, José Luis Macías2, Paul W. Layer3, Giovanni Sosa-Ceballos2, Martha Gabriela Gómez-Vasconcelos4, Guillermo Cisneros-Máximo2, Juan Manuel Sánchez-Núñez5, Joan Marti6, Felipe García-Tenorio2, Héctor López-Loera7, Antonio Pola8, and Jeff Benowitz3 8 1CONACYT 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 – Instituto de Geofísica, Universidad Nacional Autónoma de México, Antigua Carretera a Pátzcuaro 8701, 58190 Morelia, Michoacán 2Instituto de Geofísica, Universidad Nacional Autónoma de México, Antigua Carretera a Pátzcuaro 8701, 58190 Morelia, Michoacán 3College of Natural Science, Mathematics and Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK 99775 4CONACYT – Instituto de Investigaciones en Ciencias de la Tierra, Universidad Michoacana de San Nicolás de Hidalgo, Francisco J. Mujica s/n, Felicitas del Río, 58040, Morelia, Michoacán 5Instituto Politécnico Nacional-CIIEMAD, Miguel Othón de Mendizábal s/n. Col. La Escalera, C.P. 07320 Del. Gustavo A. Madero, Ciudad de México, México 6Instituto de Ciencias de la Tierra Jaume Almera, CSIC, LLuis Sole Sabaris, s/n, 08028 Barcelona, Spain 7División de Geociencias Aplicadas, Instituto Potosino de Investigación Científica y Tecnológica A.C.; Camino a la Presa San José 2055, Lomas 4a Sección, C-P. 78216, San Luis Potosí S.L.P. 8Escuela Nacional de Estudios Superiores, Universidad Nacional Autónoma de México, Antigua Carretera a Pátzcuaro 8701, 58190 Morelia, Michoacán 32 Submitted to Journal of South American Earth Sciences *Corresponding author e-mail: denisavellan@gmail.com 33 34 November 1, 2019 1 36 ABSTRACT 37 The Acoculco caldera complex (ACC) is located in the eastern part of the Trans-Mexican 38 Volcanic Belt in the northern part of the State of Puebla. The complex sits at the intersection of 39 two regional fault systems with NE-SW and NW-SE orientations. ACC was built atop Cretaceous 40 limestones, the Zacatán basaltic plateau of unknown age, early Miocene domes (~12.7-10.98 41 Ma), and Pliocene lava domes (~3.9-3 Ma). Detailed field mapping and stratigraphy studies 42 complemented by 43 30 volcanic units. Based on all these data and previous studies the ACC eruptive chronology was 44 grouped in four eruptive phases: syn-caldera, early post-caldera, late post-caldera, and extra- 45 caldera. Inception of the ACC volcanism began around 2.7 Ma with the dispersion of an andesitic 46 ignimbrite followed by the collapse of the magma chamber roof as attested by the presence of a 47 lithic breccia in isolated parts of the caldera rim. The collapse produced a 18×16 km caldera 48 depression which was partly filled by the ignimbrite (total volume of ~127 km3) followed by the 49 establishment of an intracaldera lake of unknown total extension. Early post-caldera collapse 50 activity (2.6-2.1 Ma) was restricted within the caldera producing 27 km3 of lava flows and domes 51 dominantly of basaltic trachyandesite to basaltic composition. Late post-caldera collapse activity 52 (2.0-<0.016 Ma) migrated dominantly to the caldera rim and periphery emplacing 90 km3 of 53 magma as rhyolitic domes, lavas, scoria cones, and two younger ignimbrites. The 1.2 Ma 54 Encimadas ignimbrite (26 km3) was vented through the eastern margin of the caldera and 55 dispersed to the northeast, and the 0.6-0.8 Ma Tecoloquillo ignimbrite and dome (11 km3) 56 erupted from the southwestern margin of the caldera. The most recent eruption of this phase was 57 vented close to the southeastern caldera rim producing the Cuatzitzingo (<16,710 ± 50 years BP) 58 scoria cone. Extra-caldera activity (2.4-0.19 Ma) of the Apan–Tezontepec volcanic field 40Ar/39Ar and 14C dating allowed to divide the ACC volcanic succession into 2 59 produced scoria cones and lava flows of basaltic trachyandesite to basaltic andesite composition 60 that are interbedded with the products of the caldera complex. 61 Aeromagnetic data further constraint the edge of the caldera rim and is consistent with the 62 presence of at least four intrusive bodies at depths of >1 km hosted in the Cretaceous limestones. 63 These bodies might represent a series of horizontal mafic intrusions located at different depths 64 that provide the energy that maintains active the Acoculco geothermal system. 65 66 Keywords: Geology; geochronology; geothermal energy; Acoculco caldera; Puebla 67 68 INTRODUCTION 69 Silicic collapse calderas are volcanic depressions resulting from the subsidence of the 70 magma chamber roof caused by the rapid withdrawal of magma during an explosive eruption 71 (Lipman, 2000; Martí et al., 2008; Geyer and Martí, 2014). The formation of a collapsed caldera 72 is still an enigmatic geological phenomenon because of the structural complexity involved in 73 such type of volcanic eruptions. Besides, silicic caldera eruptions represent a high risk due to the 74 large amount of magma and eruption rates involved (Costa and Martí, 2016). The resulting 75 caldera depression may represent the site of important ore deposits and high enthalpy geothermal 76 reservoirs. This makes the study of caldera systems of great interest to modern societies. In many 77 active volcanic areas, such as Iceland, New Zealand, Costa Rica, Japan, Indonesia, among several 78 others, an important part of their government energy requirements is covered by the exploitation 79 of geothermal resources associated with collapse caldera systems. In México, geothermal energy 80 has been used since 1959, when a program led by the Comisión Federal de Electricidad (CFE) 3 81 started to explore and exploit some of these potential energy resources, most of them located in 82 calderas of the Trans-Mexican Volcanic Belt (TMVB) (Hiriart et al., 2011). In the particular case 83 of geothermal fields installed in calderas, it is very important to determine the mechanism that 84 formed the caldera and the post-collapse internal structure (Bibby et al., 1995; Di Napoli et al., 85 2011; Molina et al., 2014; Afanasyev et al., 2015) to better understand fluid paths and the 86 location of reservoirs. It is crucial to determine the exact structural limits of the caldera (e.g., 87 Molina et al., 2014), the stratigraphy and age of the caldera-forming deposits and the distribution 88 of thermal anomalies. In this respect, the Acoculco caldera located in the eastern TMVB (Fig. 1) 89 represents an excellent case scenario to study the internal structure of the edifice, its eruptive 90 chronology, and surface manifestations of geothermal activity concerning fractures and faults. 91 Acoculco has been considered for years a dry-hot rock reservoir by CFE (Lorenzo-Pulido et al., 92 2010) and is the site of an on-going European-Mexican effort to develop geothermal energy from 93 non-conventional sources called GeMex Project (Calcagno et al., 2018). 94 In this contribution, we analyzed previous studies combined with new fieldwork to 95 present a simplified volcanological map and a new composite stratigraphic column of Acoculco. 96 The results are assisted by 24 new 97 units that combined with previous information set a refined evolution model of the caldera. This 98 information is crucial to present the chronology of the caldera volcanic complex through time. To 99 gain new insights on the caldera structure we produced aeromagnetic models of the area, 100 analyzed the morphostructural distribution of the volcanic units, and their correlation with 101 subsurface units in geothermal wells. Base on this new information, we propose that the late 102 Pleistocene shallow intrusions are still the heat source of geothermal activity beneath Acoculco. 40Ar/39Ar ages and one 103 104 GEOLOGICAL SETTING 4 14C radiometric date of the volcanic 105 The Acoculco Caldera Complex (ACC) is located within the TMVB, a calc-alkaline 106 volcanic arc produced by the subduction of the Cocos and Rivera plates beneath the North 107 American plate at the Middle American Trench (Pardo and Suárez, 1995) (Fig. 1). The Acoculco 108 is located in the eastern sector of the TMVB (Pasquaré et al., 1991). Acoculco occurs ~140 km 109 northeast of Popocatépetl volcano which defines in this region the active front of the TMVB 110 (Siebe et al., 1995; Macías et al., 2012) (Fig. 1). Acoculco is located over a 45-50 km-thick 111 continental crust (Urrutia and Flores-Ruíz, 1996), and sits at 400 km from the trench where the 112 Cocos plate plunges into the mantle (Pérez-Campos et al., 2008). This area is under an NW-SE 113 oriented extensional regime, as deduced from the alignment of volcanic vents, dike orientations, 114 extension fractures, and kinematics of faults at the Apan-Acoculco region (García-Palomo et al., 115 2002; 2018). The Acoculco caldera is bounded to the west by the Apan-Tezontepec volcanic field 116 (ATVF) (~3–0.2 Ma) (García-Palomo et al., 2002; García-Tovar et al., 2015), to the north and to 117 the east by Cretaceous limestones of the Sierra Madre Occidental (Avellán et al., 2018), and to 118 the south by early Miocene volcanics of the TMVB (García-Palomo et al., 2002). 119 Prior to the 80´s the Acoculco geology has not been extensively studied (e.g., Ledezma- 120 Guerrero, 1987; Castro-García and Cordoba, 1994), in fact, CFE produced the first regional map 121 of the area (De la Cruz-Martinez and Castillo-Hernández, 1986) followed by other detailed 122 studies as rock dating (López-Hernández and Castillo-Hernández, 1997; López-Hernández and 123 Martínez, 1996). Later geophysical studies attempted to understand the internal structure of the 124 caldera (Alatriste-Vilchis et al., 2005; Campos-Enriquez et al., 2003) and the hydrological 125 conditions (Huizar-Álvarez et al., 1997) of the Apan region. The first volcanological study of the 126 region was carried out by López-Hernández et al. (2009) who concluded that Acoculco was an 127 18-km wide caldera active from ~1.7 to ~0.2 Ma. These authors considered that Acoculco was 128 nested within the largest 32-km wide Tulancingo Caldera, which was active between ~3.0 and 5 129 ~2.7 Ma. Recently, Avellán et al. (2018) presented the first detailed geologic map and 130 stratigraphy of the caldera supported by nine 131 caldera has a semi-circular shape (18–16 km) and that was active from 2.7 to 0.06 Ma, thus, not 132 corresponding with the previous timing of the caldera formation (López-Hernández et al., 2009). 133 This geology of Acoculco was used to present a simplified geologic map of the caldera to 134 describe its geochemical evolution (Sosa-Ceballos et al., 2018) and a preliminary 3D model of 135 the Acoculco subsurface structure (Calcagno et al., 2018). 136 METHODS 137 40Ar/39Ar dates. These authors concluded that the Twenty-four whole-rock samples were crushed, sieved and washed in deionized water for 138 40Ar/39Ar 139 rock) were separated for dating. For sample Ac11, both a plagioclase mineral separate and a 140 whole rock sample were separated and dated, and for Ac69, two different rock types were 141 analyzed. The samples were irradiated in position 5C at the McMaster University Nuclear 142 Reactor in Hamilton, Canada, for 0.75 MWh. The standard mineral TCR-2 with an age of 28.619 143 Ma (Renne et al., 2010) was used to calculate the irradiation parameter, “J”. Samples Ac39 and 144 Ac123, were irradiated with the standard mineral MMHb-1 with an age of 523.5. The standards 145 were fused and except for AC-90. The samples step-heated using a laser dating system consisting 146 of 6W argon-ion laser at the Geophysical Institute, University of Alaska Fairbanks, following the 147 technique described in Layer (2000) and Layer et al. (2009). For Sample Ac90, an obsidian, 7 148 small shards were fused, with 6 of the 7 yielding enough gas to calculate a fusion age. The 149 samples were analyzed in a VG3600 mass spectrometer and the measured Ar isotopes were 150 corrected for system blank, mass discrimination and Ca, K, and Cl interference reactions, 151 according to procedures outlined in McDougall and Harrison (1999) and using the standards 152 reported in Renne et al. (2010). System blanks were 2x10-16 mol isotopic dating (Table 1). For most samples, phenocrysts-free groundmass chips (whole 6 40Ar and 2x10-18 mol 36Ar, 153 which are 5 to 50 times smaller than fraction volumes. Mass discrimination was monitored by 154 running calibrated air shots. 155 The structural analysis of the caldera system comprised a review of previous and new 156 structural data, after the evaluation of digital elevation models, aerial photographs, topographic 157 maps, fieldwork and the morphological evaluation of the landforms. The fault geometry was 158 characterized by field measurements and geomorphic analyses of the faults using ArcGIS on a 15 159 m-resolution digital elevation model from INEGI (Instituto Nacional de Estadística y Geografía). 160 Grid references on maps are in the WGS 1984 UTM Zone 14N projection. Rozeta 2.0 software 161 was used for plotting the fault data on stereographic projections (lower hemisphere) and to 162 perform trend-frequency analysis with rose diagrams. Stereonet 9 software was used for plotting 163 the fault planes and poles, from where the compression and tension areas were inferred into the 164 synthetic right dihedral diagram from the two pairs of conjugate fault systems, according to their 165 slip models (e.g., De Vicente et al., 1992). 166 To estimate the volume of the volcanic units we used the geological map, the Digital 167 Elevation Model (DEM), Spot-6 satellite images (1.5-m panchromatic and 6-m multispectral), 168 and the shaded relief DEM (15-m resolution) in the ArcMap 9.3 software. The difference 169 between the actual topography and the geomorphologic element of each unit was used to obtain a 170 z value, and to create a 3D surface geology on the shaded relief DEM and the volume with the 171 Surface Difference tool in ArcMap 9.3. 172 VOLCANIC STRATIGRAPHY 173 In this study we analyzed the regional (García-Palomo et al., 2002; López-Hernández et 174 al., 2009) and local geologic maps of the ACC and its surroundings (Avellán et al., 2018; Sosa- 175 Ceballos et al., 2018; Calcagno et al., 2018), and the geochronological data include in them. We 176 compiled a new simplified geologic map (Fig. 2) supported by 24 new 40Ar/39Ar and one 14C date 7 177 (Tables 1 and 2). The map includes three pre-Acoculco units, two pre-caldera, 30 ACC, and nine 178 extra-caldera units. The ACC units were subdivided in syn-caldera, early post-caldera, and late 179 post-caldera, corresponding to different phases of the caldera formation defined by their 180 distribution, stratigraphic position, age, mineral and chemical composition (Fig. 3). The chemical 181 variations of the ACC rocks have been previously documented and discussed and will not be 182 treated in this contribution (Sosa-Ceballos et al., 2018). Next, we succinctly described all units 183 that span in age from ~2.7 Ma to <0.016 Ma with their location given with respect to the caldera 184 rim. 185 Undifferentiated Cretaceous Limestones (Ksl) 186 The oldest rocks in the region are Cretaceous limestones (Ksl) of the Sierra Madre 187 Oriental that are exposed to the northeastern, eastern, and southeastern parts of the mapped area 188 (Fig. 2). Good outcrops appear in the Tenexapa and Ajajalpan canyons, close to the towns of 189 Chignahuapan and Zacatlán. At the Chignahuapan hot springs these rocks occur as light-gray 190 parallel stratified limestones with chert concretions sometimes affected by vertical fractures filled 191 with hydrothermal minerals (e.g., calcite). Although these rocks are not exposed inside the ACC, 192 they were cut in the two geothermal CFE wells at depths of ~1,200 m in EAC-1 (López- 193 Hernández et al., 2009), and 350-450 m in EAC-2 (Viggiano-Guerra et al., 2011). According to 194 these authors, Ksl was intruded by a light-gray phaneritic granite found at the bottom of EAC-1 195 drill hole at depths from 1800 to 2000 m. The intrusion produced a methamorphic aureole of 196 skarns (Viggiano-Guerra et al., 2011). A sample analyzed of this well resulted to be an aplite 197 made of alkali feldspar, plagioclase, quartz, amphibole, chlorite, and Fe-Ti oxides. An isochron 198 40Ar/39Ar 199 which may correspond to either younger intrusions beneath Acoculco or a reset age of an older 200 regional plutonic body (López-Hernández et al., 2009; Calcagno et al., 2018). age of this sample yielded an age of 183 ± 36 ka (Table 1, isochron age; Fig. 4A), 8 201 Zacatlán basaltic plateau (Za) 202 Za is a dark-gray, up to 200 m thick basaltic lava plateau that discordantly overlies the 203 Cretaceous limestones near the Chignahuapan and Zacatlán towns (Fig. 2). The lava flow is 204 aphanitic with columnar jointing and spheroidal weathering. The age of this unit is unknown, 205 however, its stratigraphic position indicates that Za is younger than the Cretaceous limestones. 206 Peñuela dacitic dome complex (~13 to 10 Ma) (Mv) 207 These rocks are exposed southern the ACC and belong to the first beginnings of regional 208 volcanism of the eastern part of the TMVB (Fig. 2). The oldest pre-caldera rocks are the Peñuela 209 dacitic (Pdd unit) and Quexnol andesitic dome complexes exposed to the SW and SE parts of the 210 ACC, respectively (Fig. 3). The Peñuela dome was dated with the K-Ar method at 12.7 ± 0.6 Ma 211 (García-Palomo et al., 2002). Another sample of this rock (Pdd) was dated here with the 212 40Ar/39Ar 213 Pre-caldera units (~3.9 – 3.0 Ma) (Pc) method at 10.98 ± 0.07 Ma (Table 1, plateau age; Fig. 4B). 214 The Puente (Pald) and Terrerillos (Tdld) lava domes are exposed to the N-NW and SW of 215 the geologic map, respectively (Figs. 2 and 3). These rocks are light-gray andesitic to dacitic lava 216 domes with greenish-gray enclaves. Pald yielded in this work an integrated age of 3,620 ± 22 ka 217 (Table 1; Fig. 4C). García-Tovar et al. (2015) reported a K-Ar age of 3.0 ± 0.4 Ma for Tdld, 218 which is consistent with its stratigraphic position. 219 220 Syn-caldera unit (Acoculco andesitic ignimbrite, ~2.7 Ma) (Sc) 221 The Acoculco andesitic ignimbrite (Aai) is a yellow to white massive deposit. It consists 222 of rounded pumice and angular to sub-angular accidental lava fragments (gray, pink and banded) 223 supported by a matrix of coarse to fine ash (Fig. 5A). Tube-banded pumice (greenish-gray to 224 white) with alkali feldspar and amphibole phenocrysts is also present. The Aai crops out in small 9 225 gullies and in some places of high topographic relief where it is greatly covered by younger 226 deposits (e.g., Hbl, Srl, Atal, Fig. 2). In the southwestern and northern parts of the caldera, Aai 227 underlies ≤40 m thick lacustrine deposits (Fig. 5B). The contact between Aai and the underlying 228 Pdd and Pald units was not observed, however, accidental fragments of these units occur in the 229 ignimbrite. At site 24, occurs a massive lithic breccia (mlBr) made of heterolithologic lavas and 230 scarce pumice (Fig. 5C-D). This breccia (5-7 m thick) contains angular accidental lavas set in a 231 coarse to fine ash matrix. Some lava fragments are aphanitic (light-gray and ochre), and other 232 porphyritic (dark-gray, greenish, and ochre) with clinopyroxene and plagioclase phenocrysts. The 233 mlBr matrix contains reddish aphanitic lithics and isolated crystals of plagioclase, pyroxene and 234 amphibole, disseminated pumice and silty minerals. Laterally, this mlBr grades to a massive fine 235 ash layer enriched in pumice and crystals of plagioclase and amphibole. In three sites (24, 65, and 236 114) mlBr is interbedded between flow units of Aai (Fig. 5D). 237 The Aai unit was transected at the EAC-1 exploratory well at depths between 210 and 560 238 m, where it covers the Pald unit (Pre-caldera units) and is overlain by the Pedernal rhyolitic lava 239 unit (Pdl, which is a late post-caldera unit) (Fig. 2). Well EAC-1 shows nearly 800 m of volcanic 240 materials resting atop a skarn which was interpreted as the metamorphosed calcareous basement 241 (López-Hernández et al., 2009; Viggiano-Guerra et al., 2011). These authors interpreted the 242 volcanic column from top to bottom as: Acoculco ignimbrite (0-130 m), Cruz Colorada dacite 243 (130-210 m), Alcholoya ignimbrite (210-580 m), and las Minas rhyodacite (580-790 m). 244 However, based on the new geologic map and revised stratigraphy of Acoculco, we consider that 245 the successions would correspond to our Pdl, Aai and Pald units (Fig. 3). The Pedernal rhyolitic 246 lava (Pdl) crops out around the CFE drill holes as a highly altered vesicular rock with feldspars, 247 plagioclase, quartz and mafic unrecognizable minerals and recrystallized lithic clasts. We 248 consider that these highly altered lavas were mistaken with pumice blocks and erroneously 10 249 described as the Acoculco ignimbrite. Also, we propose that their Alcholoya ignimbrite (~2.7 250 Ma; López-Hernández et al., 2009) described between 210 and 580 m depth would correspond to 251 our Acoculco ignimbrite Aai (Fig. 5A). Unfortunately, Aai was not recognized in the succession 252 of well EAC-2. This well only shows 340 m of volcanic infill (Pdl and Pald units) resting over 253 the skarn including 200 m of the porphyritic unit (Viggiano-Guerra et al., 2011) or our Pdl unit. 254 As mentioned before, EAC-1 and EAC-2 wells are 500 m apart but their stratigraphy suggests 255 that they are separated by a fault that is not recognizable at surface. By averaging these 256 thicknesses in the EAC-1 exploratory well, gullies and outcrops described at the surface (~470 m) 257 we estimate a minimum Aai volume of 127 km3. 258 A pumice fragment separated from Aai was dated with the 40Ar/39Ar method in 259 plagioclase at 2,732 ±185 ka (Avellán et al., 2018). In this work, we obtained two new whole- 260 rock 40Ar/39Ar dates of Aai pumice samples that yielded younger ages of 2,041 ± 38 and 2,185 ± 261 65 ka (Table 1, plateau age; Figs. 4D-E). However, Aai underlies several early-post caldera lava 262 flows as Aguila (Atal, 2.44 Ma), and Manzanito (Mtal, 2.2 Ma), and the Sayula dome (Srl, ~2.55 263 Ma) (see figures 2 and 3). Based on these stratigraphic relationships, we concluded that the age of 264 Aai must be older than 2.55 Ma for which the 40Ar/39Ar in plagioclase (2,732 ±185 ka) is the best 265 age approximation of the caldera collapse. 266 A sequence of ≥ 40 m thick lacustrine sediments (ls) is exposed inside the caldera rim in 267 the south, southwest, and northern parts. It consists of a tilted alternation of white clayed laminae, 268 and dark-gray cm-thick, volcaniclastic beds. These beds are made of rounded lava fragments set 269 in a fine-grained matrix barren of fossils. At sites 69 and 119, ls overlies the Acoculco ignimbrite 270 (Aai) (Figs. 2 and 3). 271 272 11 273 Early post-caldera units (~2.6-~2.1 Ma) (Epc) 274 Four of these units are exposed inside the caldera depression as basaltic to basaltic 275 trachyandesite lava flows (Hbl, Atal, Vtal and Mtal) that partially cover Aai (Figs. 2 and 3). They 276 are highly eroded and have asymmetric morphologies similar to flatirons with their apex towards 277 the center of the caldera developing a sub-radial exorheic drainage. These units typically appear 278 as light-gray to dark-gray, blocky lava flows with porphyritic to aphanitic textures. These lavas 279 frequently present greenish, yellowish and reddish hydrothermal alteration zones, and host 280 xenoliths of sub-rounded limestones, sandstones, and fine grain granite. Two units were dated 281 with the 40Ar/39Ar method at 2,323 ± 48 ka (Vtal) and 2,199 ± 24 ka (Mtal) (Table 1, plateau age; 282 Figs. 4F-G). Avellán et al. (2018) obtained a 283 Atal. The age of Hbl is unknown; nevertheless, it underlies the Srl dome dated at ~2.55 Ma (see 284 below). The other two early-post-caldera units (Srl and Atad) are exposed to the external 285 northwestern and southeastern rim of the caldera, respectively. The Srl unit (2,553 ± 110 ka; 286 Table 1; Fig. 4H) consists of gray to black and brown, banded obsidian lava flows with 287 holohyaline texture and partially devitrified to light-gray spherulites and lithophysae. The Atad 288 unit (2,179 ± 26 ka; Table 1; Fig. 4I) is a blocky greenish to light-gray, porphyritic lava flow. For 289 all the early post-caldera units we estimated a volume of 27 km3. 40Ar-39Ar whole-rock age of 2,441 ± 234 ka for 290 291 Late post-caldera units (~2.0 - ~0.016 Ma) (Lpc) 292 These units are represented by 11 domes (Alrd, Lrd, Amrc, Prld, Crcd, Trcd, Crd, Ard, 293 Crl, Arcd and Mrcd), 5 lava flows (Coal, Tal, Cual, Pdl and Prld), 2 ignimbrites (Eri, Tr1), and 4 294 scoria cones (Plc, Tlc, Clc1 and Clc2). The Alrd, Lrd, Amrc, Crcd, Trcd, Crd, Ard, Crl Arcd and 295 Mrcd dome and lava flow units occur on the caldera border and periphery (Fig. 2). These 296 structures have predominant rhyolitic compositions with coulée and asymmetrical morphology 12 297 delimited by very steep levées. The rocks of these units typically appear as light-gray to pinkish- 298 gray, banded to massive obsidian lavas with mottled structure given by abundant spherulites and 299 lithophysae (Fig. 3). López-Hernández et al. (2009) and García-Tovar et al. (2015) reported K-Ar 300 ages in hornblende and whole-rock of the Lrd (1,700 ± 400 ka), Crd (1,300 ± 600 ka), and Crl 301 (1,274 ± 27 ka). In this work, seven new units were dated with the 302 yielding ages of 1,870 ± 36 ka (Alrd) (Fig. 4J), 1,438 ± 24 ka (Amrc) (Fig. 4K), 1,394 ± 8 ka 303 (Crcd) (Fig. 4L), 1,360 ± 15 ka (Trcd) (Fig. 6A), 1,283 ± 88 ka (Ard) (Fig. 6B), 998 ± 36 to 304 1,145 ± 14 ka (Arcd) (Fig. 6C-D) and 1,066 ± 42 ka (Mrcd) (Fig. 6E), that are in agreement with 305 their stratigraphic position. On the other hand, on the western edge the undifferentiated Maguey 306 unit formed of a sequence of pyroclastic surges and fall with a 307 (Table 1; Fig. 6F). 40Ar/39Ar 40Ar-39Ar method (Table 1) age of 1,084 ± 22 ka 308 The Coal, Tal, Cual, Pdl and Prld units are situated inside the caldera partially overlying 309 the syn-caldera and early post-caldera units (Fig. 2). They appear as a stack of bedded lava flows 310 with andesitic (Coal, Tal, and Cual) and rhyolitic (Pdl and Prld) compositions. The andesitic lava 311 flows are porphyritic, black to dark-gray in color, with reddish to yellowish intense hydrothermal 312 alteration (Fig. 3). The Coal unit was dated by Avellán et al. (2018) with the 40Ar-39Ar method at 313 2,027 ± 40 ka. A dike 40Ar-39Ar age that cuts this unit yielded 1,600 ± 35 ka (Table 1; Fig. 6G). 314 In this study we dated the Tal unit with the same method at 1,708 ± 54 ka (Table 1; Fig. 6H). 315 López-Hernández et al. (2009) obtained a 40Ar/39Ar age of 1,600 ± 200 ka for the Cual unit. 316 The rhyolitic lava flows (Pdl and Plrd units) are pinkish-gray to pinkish-white porphyritic 317 rocks. These units are white and highly hydrothermally altered lavas that crop out in the vicinity 318 of Pedernal and Acoculco towns. The lava flows are corroded by hydrothermal fluids along 319 highly vesicular breccia structures, where phenocrysts and matrix have been replaced by 13 320 alteration minerals. López-Hernández et al. (2009) reported two K-Ar ages for these lavas of 321 1,600 ± 100 ka (Pdl), and 1,400 ± 200 ka (Prld). 322 The Encimadas unit (Eri) is a rhyolitic ignimbrite widely exposed on the east-northeast 323 external parts of the ACC. It has a moderately dissected plain that mantles the Zacatlán basaltic 324 plateau and partially covers some of the post-caldera units. Eri is a welded ignimbrite with 325 several flow units that appear as massive, light-gray to white, beds. Each bed consists of matrix- 326 supported fine ash particles with feldspar and quartz phenocrysts. It has an approximated volume 327 of 26 km3. López-Hernández et al. (2009) reported a 40Ar/39Ar age in sanidine of Eri at 1,300 ± 328 200 ka. A similar 40Ar/39Ar age of 1,278 ± 14 ka was obtained in this study (Table 1, plateau age; 329 Fig. 6I). 330 The Tecoloquillo ignimbrite (Tr1) is a rhyolitic ignimbrite widely exposed to the south- 331 southwest parts of the caldera (Fig. 2). South of the caldera, Tr1 partially covers some units 332 belonging to the pre-caldera, syn-caldera and post-caldera. The Tr1 unit consists of two main 333 beds; the lowermost part is massive, monolithologic, brittle and matrix-supported with highly 334 friable pumice fragments embedded in a medium to fine ash matrix. Both pumice and matrix 335 contain bipyramidal quartz and alkali feldspar phenocrysts. The upper part is massive with pink- 336 gray, corroded lava blocks, supported by a crumbly medium ash matrix. The upper part of Tr2 is 337 a rhyolitic dome made of angular light-pink to gray lava blocks. The rock is moderately vesicular, 338 fibrous and porphyritic with quartz, alkali feldspar and amphibole phenocrysts. One pumice 339 sample collected from the basal part of the Tr1 unit yielded a 40Ar/39Ar age in plagioclase of 611 340 ± 72 ka (Table 1; Figs. 6J). Other pumice sample that yielded older age of 762 ± 9 ka (Table 1; 341 Fig. 6K). López-Hernández et al. (2009) reported a 40Ar/39Ar age in sanidine of 0.8 ± 0.1 Ma for 342 this unit. 14 343 The last four late-post-caldera units are cinder cones (Paila, Tulimán, Cuatzitzinguito and 344 Cuatzitzingo) and associated lava flows of basaltic andesite composition. These scoria cones 345 occur above or close to the topographic caldera rim (Fig. 2). The Paila (Plc) and Tulimán (Tlc) 346 cinder cones are exposed on the southeastern and northwestern parts of the caldera rim, 347 respectively. The Paila unit directly overlies the Atad and Eri post-caldera units and Tulimán unit 348 lies discordantly on top of Srl post-caldera unit. The Cuatzitzinguito (Clc1) and Cuatzitzingo 349 (Clc2) scoria cones lies on the southern flank of the Paila unit (Fig. 3). All these scoria cones are 350 composed of massive poorly-sorted fallout beds with dense blocks and bread-crust scoria and 351 spatter bombs, as well as, black to dark-gray blocky lava flows associated with effusive activity. 352 The Tulimán scoria cone was dated with the 40Ar/39Ar at 63 ± 9 ka (Avellán et al., 2018). A 40Ar- 353 39Ar 354 underneath a fallout tephra of the Cuatzitzinguito scoria cone yielded a 355 BP (Table 2). We did not obtain an age for the Cuatzitzingo scoria cone, however, its 356 stratigraphic position indicates that it should be younger than 16 ka. age of 71 ± 17 ka was obtained here for the Paila scoria cone (Table 1; Fig. 6L). A paleosol 14C age of 16,710 ± 50 We estimated a total volume for all the late-post caldera units including Eri and Trl of 90 357 358 km3. 359 Extra-caldera units (~2.4-0.19 Ma) (ATVF) 360 Nineteen scoria cones and four small-shield volcanoes of the ATVF occur around the 361 caldera complex (Figs. 2). Four of these scoria cones, known as Amanalco (Asc; 2,408 ± 58 ka; 362 Avellan et al., 2018), Huixtepec (Hsc), Tecolote (Tsc) and Apapasco (Asc) are located ca. 7 km to 363 the southeast of the caldera border. Five scoria cones called Buenavista (Bsc), Comal (Csc), 364 Calandria (Csc), Toronjil (Tsc), and Tezontle (Tsc) and the Coatzetzengo small-shield-volcano 365 are located at ca. 4 km to the northwest. Three scoria cones named Moxhuite (Msc; 239 ± 34 ka; 366 Avellán et al., 2018), Matlahuacala (Msc), and Cazares (Csc) lie discordantly on top of the 15 367 Encimadas ignimbrite at ca. 6 km to the east. Two small-shield volcanoes, Camelia (Clc; 2,033 ± 368 84 ka; Avellán et al., 2018) and Tetelas (Tlc, 1,060 ± 84 ka; Avellán et al., 2018) are located at 10 369 km to the south of the caldera border. Three scoria cones are aligned in a NW-SE direction, these 370 are Tecajete (Tsc) (1,235 ± 62 ka), Blanco (Bsc; 1,274 ±62 ka; Avellán et al., 2018) and Hermosa 371 (Hsc). Another, four scoria cones, Coliuca (Clc; 188 ± 6 ka K-Ar age; García-Tovar et al., 2015), 372 Colorado (Csc), El Conejo (Csc) and Tezoyo (Tsc), and the Coyote small-shield volcano are 373 situated at ca. 7 km to the southwest of the caldera border. 374 CHARACTERIZATION OF FAULT SYSTEMS 375 The ACC is in a highly tectonized region affected by NW-SE, NE-SW and E-W 376 structures (Fig. 7A). The NW-SE and NE-SW structures belong to two regional fault systems. 377 The NW-striking (NNW to NW) fault system contains the oldest regional structures. Some of 378 these faults follow the trend of older structures, such as fold axes and thrust faults from the 379 Laramide orogeny in the Sierra Madre Oriental (Suter, 1984; Rocha et al., 2006; Lermo et al., 380 2009), exposed just north of the study area. The NE-oriented Cenozoic extension (Henry and 381 Aranda-Gómez, 1992) was active in this region from middle to late Miocene creating NW- 382 striking normal faults and NE-striking strike-slip faults synchronous with regional volcanism 383 (Andreani et al., 2008). In the ACC, the NW-SE fault system is represented by several major 384 normal and oblique (right-lateral component) faults and numerous minor fault strands with 385 lengths between 2 and 5 km, with an average azimuth of 130˚ dipping mainly to the NE at 386 average angles of 50˚, creating horst and graben-like structures (García-Palomo et al., 2002; Fig. 387 7A). The Manzanito structure, also known as the Tulancingo-Tlaxco fault system (López- 388 Hernández et al., 2009), is the most representative fault pertaining to this system in the ACC 389 (Avellán et al., 2018). It has an en échelon array of normal faults, which extends for ca. 30 km 390 (Fig. 6B) and displaces the 1.7 Ma Lobera rhyolitic dome by at least 145 m, and the 1.07 Ma 16 391 Minilla rhyolitic dome by 120 m (this study). Also, at around 1.6 Ma, the ~160˚-striking and ~2.5 392 km long Colorada basaltic dikes were emplaced at the southeastern part of the caldera (Fig. 2; 393 Coal; Avellán et al., 2018). But there is no evidence of recent (Holocene) activity in the local 394 NW-striking faults. 395 The NE-striking (NNE to NE) Apan-Tlaloc Fault System (García-Palomo et al., 2018), 396 Tenochtitlan-Apan Fault System (García-Palomo et al., 2002) or Tenochtitlan Shear Zone (De 397 Cserna et al., 1988), has been active since the Miocene and is still active (García-Palomo et al., 398 2002; 2018; Fig.7). This is the most important fault system in the region, consisting of normal 399 faults with a left-lateral component with an average azimuth of 040˚ dipping both to the NW and 400 SE with an average dip angle of 75˚ that present two generations of striae (~30 and 80°). This 401 system has created regional horst and graben-like structures, obeying a NW-oriented extension 402 regime that affects eastern México (this study and García-Palomo et al., 2018; Fig. 7A). In 403 general, this fault system shows a good geomorphic expression represented by several major 404 faults (e.g. Apan-Tlaloc Fault and Chignahuapan Fault) and numerous minor fault strands and 405 fractures with 1-4 km individual lengths. Field exposures show prominent scarps; the ~1.3 Ma 406 Canoas rhyolitic dome is displaced by the Atotonilco fault (northern caldera rim) by 150 m, and 407 the ~2.2 Ma Ajolotla trachyandesitic dome is displaced by the Chignahuapan fault by 200 m (this 408 study). The most recent volcanic structures of the ACC (Plc with 71 ka, and Tlc with 63 ka units) 409 are cut by these faults (this study; Figs. 2 and 7B). 410 The NE- and NW-striking normal fault systems intersect each other (García-Palomo et al., 411 2002; Lermo et al., 2009) creating an orthogonal arrangement of grabens, half-grabens and 412 horsts. The NE-SW Rosario-Acoculco Horst (García-Palomo et al., 2002) is delimited to the west 413 by the 235°-trending and NW-dipping Apan-Tlaloc Fault (Mooser and Ramírez, 1987; Huizar- 414 Álvarez et al., 1997) and to the east by the 055°-trending and SE-dipping Chignahuapan Fault 17 415 (Avellán et al., 2018). The ACC is a volcano-tectonic depression inside the Rosario-Acoculco 416 Horst, delimited by parallel faults but with opposite dipping. As for the E-W structures, they are 417 represented by at least 10 E-W striking (~N085°) and SE-dipping normal fault strands. They can 418 only be observed inside the ACC, mainly affecting post-caldera volcanism in the central part of 419 the ACC (this study; Figs. 2 and 7). 420 AEROMAGNETIC DATA 421 We produced an aeromagnetic map by reprocessing the airborne data of the Mexican 422 Geological Survey obtained in 2000 (Fig. 8A). This map shows different aeromagnetic anomalies 423 inside or outside of the ACC. The outer domains may be caused by regional and topographic 424 anomalies. Instead, inner domains are possibly associated to four positive local anomalies 425 (subdomains) and conforming a semi-circular shape of the Acoculco caldera. The first subdomain 426 is located in the central part of the ACC; it shows an NE-oriented elongated shape (9.7 km long 427 and 4.8 km wide) with magnetic intensities between -81.8 and 57.8 nT. The second anomaly is 428 located at the southeastern boundary of the ACC, ENE-oriented, 6.6 km long and 4.1 km wide 429 with magnetic intensities between -135.1 and 25.7 nT. The third anomaly is located in the south- 430 central portion of the ACC, NE-oriented, 8.9 km long and 3.5 km wide with magnetic intensities 431 between -64.5 and 36.5 nT. The fourth anomaly appears in the western part of the ACC, NW- 432 oriented, 5.4-km long and 3-km wide with magnetic intensities between 84.5, and -2 nT. The map 433 shows a contrast between the caldera (higher magnetic values) and the surrounding areas (lower 434 magnetic values). 435 DISCUSSION 436 Tectonic implications 18 437 The regional tectonic setting has controlled the formation and evolution of the ACC. The 438 ACC formed on top of the Rosario-Acoculco horst, which is bounded by the Apan and Tlaxco- 439 Chignahuapan grabens (Fig.7). Furthermore, some pre-existing tectonic faults were used as 440 pathways for dike intrusions and volcanism (aligned scoria cones parallel to the NE-trending fault 441 system), but also as part of the ring fault system that controlled the caldera collapse and exerted 442 the main control on the location of post-caldera vents (Fig. 2). The southwestern caldera rim 443 coincides with the NW-SE Manzanito fault, while the northern caldera rim is marked by the 444 Atotonilco scarp (Avellán et al., 2018). The sub-linear to sub-circular collapsed structures form a 445 18 x 16 km rhombohedral-shape, dipping towards the central part of the ACC. In fact, the left- 446 lateral movement in the NE-SW faults during the middle Miocene (García-Palomo et al., 2000; 447 2018) and/or the right-lateral movement of the NW-SE faults could have been responsible for the 448 fracturing and creation of the space necessary to accommodate the magma reservoir beneath the 449 caldera (pull-apart basins in transtensional regimes; e.g., Bursik, 2009; Saxby et al., 2016). 450 Activity of the NE-SW and NW-SE regional systems continued after the caldera formation and 451 modified the trace of the caldera border, causing its displacement at several points (e.g., 452 northeastern tip of the topographic caldera rim, outcrops 27 and 28; Fig. 7B). Tectonic 453 movements affected the interior of the caldera until very recent time, causing displacements of 454 intra-caldera blocks and disturbing the position of intra-caldera volcanic vents and products. For 455 example, E-W faulting in the central part of the caldera (outcrops 24, 32 and 54; Fig. 7B) 456 Caldera reactivation during the emission of the 26 km3 Encimadas (~1.28 Ma) and 11 km3 457 of the Tecoloquillo ignimbrites (~0.6-0.76 Ma) was possibly originated near or along the E and 458 SW borders of the caldera, respectively. Therefore, it would not be surprising if the caldera used 459 these faults systems to nucleate the rest of the ring fault that controlled the collapse event, as it 460 has occurred in other well documented calderas (see Aguirre-Díaz et al., 2008; Martí et al., 2013; 19 461 Molina et al., 2014). Despite the strong, selective erosion that has affected parts of the area, the 462 northern topographic caldera rim has not retreated significantly from its original position. This 463 may imply that caldera subsidence continued for a long time (±500 ka, between Encimadas and 464 Tecoloquillo eruptive events), dissecting younger rocks emplaced close or through the ring faults. 465 On the contrary, in the eastern side of the ACC, neither the morphological nor the structural 466 border are visible at surface, however, it is depicted in the aeromagnetic map (Fig. 8A). At the 467 surface, the eastern border is not visible because it has been buried by younger deposits and it is 468 also likely an uneven collapse of the caldera. The E-W intra-caldera fault system is limited to the 469 caldera interior and therefore is could represent local deformation related to the resurgence of the 470 caldera. 471 The NE-SW and NW-SE fault systems coexist in the ACC (Fig. 7B). This can be 472 explained by two different constructive phases. The first one took place in the Miocene, ruled by 473 ~NE-oriented extension, forming ~NW-SE dip-slip faults and ~NE-SW sinistral strike-slip faults 474 (Henry and Aranda-Gómez, 1992). The second phase took place in the Pliocene-Pleistocene, and 475 it is controlled by ~NW-oriented tectonic extension, forming ~NW-SE dextral strike-slip faults 476 and ~NE-SW dip-slip faults (oblique and normal faults with a minor left-lateral component; 477 García-Palomo et al., 2018). This is consistent with the left-stepping en échelon geometry of the 478 NW-SE Manzanito fault originated by right-lateral slip. Actually, this latter phase of NW- 479 oriented extension is still active (García-Palomo et al., 2018). Therefore, both fault systems occur 480 in the same region and are still active under the same stress regime. This is possible because the 481 NW-SE structures are acting as transfer faults of the NE-SW normal faults. In order to 482 demonstrate that both fault systems are moving under the same strain field, a synthetic right 483 dihedra diagram was created for the two pairs of conjugate fault sets, according to the mean 484 measured planes (Fig. 7B). This slip model implies an orthorhombic symmetry of faults, where 20 485 superposed areas divide the figure into the different compressional and tensional areas in the 486 ACC. 487 Aeromagnetic interpretation and geothermal implications 488 Previous studies used the aeromagnetic and gravimetric data of the Acoculco-Atotonilco 489 region obtained by Petroleos Mexicanos (PEMEX) in 1980 (García-Estrada, 2000; López- 490 Hernández et al., 2009). These authors interpreted a depression bounded by faults coincident with 491 gravimetric data and low magnetic values with much higher values at the interior defining the 492 border of the Acoculco caldera. This coincides with our interpretation of the airborne data 493 displayed in the aeromagnetic map of figure 8A. This map shows an approximated match 494 between the edge of the 18 x 16 km Acoculco caldera as reconstructed from the distribution of 495 the Acoculco ignimbrite, the co-ignimbritic breccia, the Atotonilco scarp, and the Manzanito fault 496 (Avellán et al, 2018; this work) (Fig. 7). The aeromagnetic anomalies suggest a rough elliptical 497 limit of the caldera that may be bounded to the east by the approximated venting location of the 498 Encimadas ignimbrite and to the southeast by the basaltic andesite Paila scoria cone. The positive 499 aeromagnetic anomaly located at the NE portion of the ACC may be produced by the Zacatlán 500 basaltic plateau that is overlaid by the Encimadas ignimbrite and both cover the Cretaceous 501 limestones (Fig. 2). The positive anomalies located to the SW and NW of the ACC may be 502 associated with intrusive bodies that gave place to monogenetic volcanoes of the ATVF. The 503 interior of the caldera is not homogeneous but includes higher and lower magnetic values that 504 might reveal the presence of small-scale horsts and grabens combined with the presence of 505 shallow intrusive bodies (i.e., laccoliths, sills) of basaltic and basaltic and intermediate 506 composition intruded inside the Cretaceous limestones at depths (>1 km; Fig. 8B). These positive 507 anomalies inside the caldera could have fed the ca. 2.6 Ma Huistongo, 2.2 Ma Manzanito, 0.71 508 Ma Paila and <16 ka Cuatzitzingo eruptions (Figs. 2, 3 and 7). The anomalies described in figure 21 509 8A have rough NE-SW orientations separating a local horst on top of which the CFE exploratory 510 wells were drilled, cutting skarns at ~790 m deep (EAC-1), and argilitic limestones at 350 m deep 511 (EAC-2) (Viggiano-Guerra et al., 2011). These positive magnetic anomalies might represent 512 intrusive bodies injected during the post-caldera phase (Huistongo lavas 2.6 Ma), which may 513 have an important role to keep the geothermal system active. Sosa-Ceballos et al. (2018) 514 concluded that after the caldera collapse some 2.7 Ma ago, the stress field on the ACC magma 515 chamber and its immediate surroundings was modified. Such change hindered magma migration 516 through the collapsed reservoir and promoted lateral tension zones where dikes and sills converge 517 and eventually form new magma chambers. Due to the active tectonics in the Acoculco area, a 518 new plumbing system developed after the caldera collapse and gradually favored the ascent of 519 deep seated, independent peralkaline and calcalkaline mafic magmas. In addition, Sosa-Ceballos 520 et al. (2018) found that magma mixing-heating is the main magmatic process that modified the 521 ACC rock suite. Hence, the aeromagnetic anomalies might represent the horizontal intrusion 522 zones where mafic magma accumulates as a consequence of the upper crustal deformation 523 produced by the collapse. These intrusion zones serve as heating elements of magma reservoirs 524 (that eventually might evolve into small magma chambers) and yield the energy that sustain the 525 Acoculco geothermal system. When magma did not migrate horizontally, it flowed upwards, 526 sometimes getting tapped on its way to the surface evolving to silicic compositions es got trapped 527 as the aplitic dike dated at 183 ± 36 ka found in EAC-1 (Table 1). These young felsic intrusions 528 have reactivated the system providing heat for the hydrothermal activity and Holocene 529 hydrothermal explosions (Canet et al., 2015a and b). López-Hernández et al. (2009) also 530 suggested the possible presence of an intrusive body, whose top is at > 1000 m depth and that 531 was not reached by EAC-1 well. Although, we believe these intrusions also contribute to the 22 532 geothermal system, their volume and abundance are not yet known and probably can be estimated 533 by geophysical methods to constraint their energy input. 534 The distribution of magnetic anomalies inside the caldera also reveals the presence of 535 tectonic lineaments with NE-SW and NW-SE orientations, likely associated to horst and grabens 536 blocks and coinciding with those intra-caldera and caldera ring faults observed and measured in 537 the field, suggesting that such faults were used as vent zones for post-caldera volcanism (Figs. 7 538 and 8). These faults may have also favored the ascent of gas with isotopic compositions (N2/He, 539 3He/4He, 13C, 15N) 540 of CO2 and 3He/4He ratio (R/Rair = 8.5), that suggest an active magmatic source at depth under 541 the ACC (Polak et al., 1982). 542 Volcanic evolution of both MORB- and arc-type signatures (Peiffer et al., 2014), and high values 543 Based in our synthetic map, refined volcanic stratigraphy, structural analysis, and 544 interpretation of the aeromagnetic anomalies ACC was emplaced through the Cretaceous 545 limestones, the Peñuela dacitic domes (~13 to 11 Ma), and the Zacatlán basaltic plateau (Figs. 2 546 and 3). Rocks of the ACC were divided into 30 units formed in different stages of the caldera 547 evolution from pre, syn, early-post, late- post caldera and extra-caldera volcanism of the Apan- 548 Tezontepec Volcanic Field (Avellán et al., 2018; Sosa-Ceballos et al., 2018). 549 Caldera formation (~2.7 Ma) 550 Prior to the formation of the caldera, the Puente and Terrerillos lava domes had been 551 extruded in the area (Fig. 9A). These domes of andesitic and dacitic composition and calc- 552 alkaline affinity were emplaced during the Pliocene between 3.9 and 3 Ma. After a few hundred 553 thousand years (0.3 Ma), a large amount of andesitic magma stagnated at depth preparing for a 554 major eruption. This calc-alkaline magma overpressured the encasing rock initiating the emission 23 555 of an andesitic pyroclastic density current (Acoculco ignimbrite). The continuous emission of the 556 ignimbrite eventually diminished the magmatic pressure and weakened the roof of the magma 557 chamber triggering its collapse forming a 18 x 16 km asymmetric caldera (Fig. 9B). The collapse 558 followed older structures as the NW-SE Manzanito fault in the western and southwestern parts 559 and generated a semi-circular caldera ring (Atotonilco scarp) at the north (Fig. 7B). The 560 ignimbrite crops out in the western-southwestern and northern parts of the caldera interior 561 because in other locations it was covered by younger deposits. The used of older and fractures as 562 magma feeders facilitated a high discharge rate at the beginning of the eruption as discussed by 563 Costa and Martí (2016). This developed into massive proportions, thus precluding the formation 564 of a vertical eruption column and the deposition of fallout deposits, which have not been 565 recognized neither outside the caldera nor in the intra-caldera wells. The occurrence of a co- 566 ignimbrite breccia found in relative small, isolated areas at the north, west and southwestern parts 567 of the caldera rim points to emissions sites of the ignimbrite. In fact, the ring fault scarps and the 568 topographic caldera rim are clearly visible at the western half of the caldera, but they are hidden 569 at the eastern side, suggest that caldera-collapse could have been more intense in the western 570 sector than in the eastern sector. We cannot inform on how the caldera-forming eruption 571 progressed, as most if its corresponding deposits have been eroded outside the caldera or are now 572 totally hidden by post-caldera rocks at the interior of the caldera depression. However, the 573 presence of a large number of normal faults with different orientations affecting intra-caldera 574 rocks suggests that caldera-collapse could have occurred in a piecemeal-trapdoor fashion through 575 several stages of volcanism. 576 Early-post caldera phase (2.6-2.2 Ma) 577 After the caldera formation followed a relatively short (~0.1 Ma) quiescence period 578 during which an intra-caldera shallow lake developed with lacustrine sedimentation (Fig. 5B). 24 579 The early post-caldera volcanism is represented by eruptions that occurred mainly in the interior 580 of the caldera with a minimum total volume of 27 km3 (Fig. 9C). This volcanism vented along the 581 ring faults and intra-caldera normal faults that formed or acted during caldera collapse to then be 582 reactivated by the extensional tectonics that have affected the area until present (García-Palomo 583 et al., 2018). These modifications of the local stress field allowed the ascent of peralkaline 584 basaltic and basaltic trachyandesite magmas that were generated by partial melting of a 585 metasomatized mantle, genetically unrelated to the calc-alkaline magmas that dominated the pre- 586 caldera and caldera activities (Sosa-Ceballos et al., 2018). Both suites of magma mixed and 587 formed the early post-collapse dome complexes on the ring of the caldera. 588 Late-post caldera phase (~2.0 - ~0.016 Ma) 589 After a quiescent period of around 0.2 Ma the caldera entered a new phase of volcanism 590 along the caldera rim dominated by the emission of domes and lava flows, two ignimbrites, and 591 four scoria cones (Fig. 9D). An extended period of intermittent volcanism occurred between 2 592 and 1.3 Ma with the occurrence of at least 13 effusive eruptions along the caldera rim. Between 2 593 and 1.6 Ma eruptions emitted bimodal andesitic (Coal, Tal, Cual units) and rhyolitic (Alrd, Lrd 594 units) products to then erupted rhyolitic effusive magmas until 1.28 Ma (Ard unit) at the 595 southwestern part of the caldera rim. At about (1.27 Ma) a rhyolitic explosion occurred at the 596 eastern part of the caldera dispersing to the east and northeast the Encimadas ignimbrite (26 km3). 597 We assume that magmatic overpressure in a shallow magma chamber triggered the rhyolitic 598 eruption that represents the second largest event of the caldera. Immediately after this major 599 eruption occurred four rhyolitic eruptions in the S-SW parts of the caldera between 1.27 and 1.06 600 Ma. These eruptions particularly emitted the voluminous Ailitla coulée dome to the south (Arcd; 601 12.4 km3), the Cabezas lava to the southwest (Crl), a minor explosive event that dispersed small- 602 volume pyroclastic density currents (Msf), and the Minilla coulée come to the west (Mrcd). These 25 603 effusive eruptions preceded the occurrence of another explosive rhyolitic event that occurred 604 some 0.2 Ma later. This eruption vented at the southwestern edge of the caldera rim (0.6-0.76 605 Ma) at the intersection with the NW-SE Manzanito fault. This eruption dispersed the 606 Tecoloquillo rhyolitic ignimbrite (11 km3) to the southwest of the caldera rim ending with the 607 extrusion of a rhyolitic summit dome. The Tecoloquillo ignimbrite represents the third largest 608 explosive event of the ACC after which activity associated to the caldera apparently declined for 609 a period of circa 0.5 Ma. However, renewed activity of the caldera started 0.071 Ma with the 610 emplacement of the La Paila (Plc) at the southeast, the Tulimán scoria cone and lava flow (0.063 611 Ma) at the northwest, and the Cuatzitzinguito (16 ka) and Cuatzitzingo ca. 0.01 Ma scoria cones 612 at the southeast. These magmas erupted basaltic andesite to basaltic trachyandesite products 613 (Sosa-Ceballos et al., 2018). 614 During this late post-caldera period the peralkaline magma suite gradually dominated 615 over the calc-alkaline suite (Sosa-Ceballos et al., 2018) (Figs. 9C-D). This process resulted in an 616 undistinguishable unique source for the early post-collapse magmas, whereas trace elements 617 geochemistry suggests a relatively homogeneous source for the late post-collapse rhyolites (Sosa- 618 Ceballos et al., 2018). The total volume of magma erupted during late post-caldera phase was 619 ~90 km3. We believe that during this phase of the caldera evolution began the resurgence of the 620 caldera floor as suggested by uplifted and tilted lacustrine sediments (250°/30SE) exposed inside 621 the caldera (Fig. 2) covering the Acoculco ignimbrite. The morphology of the caldera floor 622 suggests that the resurgence was more important in the southwestern portion, which is in 623 accordance with shallow intrusions of magma revealed by aeromagnetic data that could cause the 624 caldera resurgence. According to Sosa-Ceballos et al. (2018) these shallow magma intrusions 625 were originated by post-caldera deformation that promoted the formation of sills and dykes above 626 the collapsed reservoir and might be the heat source of the Acoculco geothermal system. 26 627 Finally, the area covered by the AAC products is ca. 856 km2 with a total minimum 628 volume ca. 143 km3. By assuming that the present extension of deposits is minimum, we 629 calculated that activity was dominated by nearly 80 % of effusive volcanism with three major 630 bursts of explosive volcanism that represent 20% of the ACC. As summarized above Acoculco 631 has been a site of persistent volcanism since 2.7 Ma that include the presence of a young 632 intrusion (183 ± 36 ka) at the bottom of the EAC-1 well. This intense magmatism, as well as 633 Holocene hydrothermal explosions (Canet et al., 2015a; 2015b), and geothermal manifestations 634 (Peiffer et al., 2014) indicate that the complex is still active and could represent a site to develop 635 a geothermal field. 636 CONCLUSIONS 637 The Acoculco collapse caldera was originated at about ~2.7 Ma ago in response to the 638 eruption of 127 km3 of the andesitic ignimbrite of the Acoculco unit. The caldera collapse episode 639 occurred in part along a pre-existing regional NE-SW trending faults (southwestern, western and 640 northwestern sectors), but also along newly formed ring faults (northern sector and probably the 641 eastern sector, but this is not visible). The collapse occurred in a piecemeal-trapdoor fashion, in 642 which intra-caldera blocks bounded by pre-existing or newly formed normal faults gravitationally 643 collapsed in a partly emptied magma chamber. After the formation of the caldera, volcanic 644 activity stopped for a while permitting the formation of an intra-caldera lake as suggested by 645 sediments covering part of the caldera-forming ignimbrite. Renewed volcanic activity emplaced 646 several domes and lava flows along the caldera borders and intra-caldera faults that modified the 647 caldera depression. Post-caldera volcanism also affected the areas around the caldera and was 648 controlled by the main regional fault systems that also influenced the formation of the caldera. 649 The ACC was formed at the intersection of NE-SW and NW-SE fault systems. These fault 27 650 systems have controlled the formation and evolution of the ACC, the regional subsidence, and 651 venting of syn- and post-caldera volcanism inside and outside the caldera. 652 The presence of positive aeromagnetic anomalies associated to intrusive bodies (sills and 653 dykes) at the interior of the Acoculco caldera makes it an interesting target for geothermal 654 exploration. 655 ACKNOWLEDGEMENTS 656 This study was partially funded by the Centro Mexicano de Innovación en Energía 657 Geotérmica (CeMIE-Geo) project P15 and GeMex 4.4. to J.L. Macías. We thank F. Mendiola, G. 658 Reyes-Agustín and S. Cardona for their technical support during the laboratory analyses and D.E. 659 Torres-Gaytan for his support during the aeromagnetic map generation. We appreciate the 660 discussions and input provided by C. Arango, C. Canet and J. Marti is grateful for the MECD 661 (PRX16/00056) grant. We appreciated the constructive comments by J. Aranda and two 662 anonymous reviewers. 663 REFERENCES CITED 664 Afanasyev, A., Costa, A., Chiodini, G., 2015. Investigation of hydrothermal activity at Campi 665 Flegrei caldera using 3D numerical simulations: Extension to high temperature 666 processes. Journal of Volcanology and Geothermal Research, 299, 68-77. 667 Aguirre-Dı́az, G.J., López-Martı́nez, M., 2001. The Amazcala caldera, Queretaro, Mexico. 668 Geology and geochronology. Journal of Volcanology and Geothermal Research, 111(1-4), 669 203-218. 670 Aguirre-Díaz, G.J., Labarthe-Hernández, G., Tristán-González, M., Nieto-Obregón, J., Gutiérrez- 671 Palomares, I., 2008. The ignimbrite flare-up and graben caldera of the Sierra Madre 672 Occidental, México. In: Gottsmann J, Martí J. (Eds.), Caldera volcanism: analysis, modeling 673 and response. Elsevier, Amsterdam, 143–174. 674 Alatriste-Vilchis, D., Campos-Enriquez, O., Huizar-Alvarez, R., Marines-Campos, R., 2005. La 675 estructura sub-superficial de la subcuenca de Tecocomulco. La Laguna de Tecocomulco Geo- 28 676 Ecologia de un desastre. Universidad Nacional Autónoma de México, Instituto de Geología, 677 Publicación especial, 3, 33-48. 678 Angelier, J., 1990. Inversion of field data in fault tectonics to obtain the regional stress—III. A 679 new rapid direct inversion method by analytical means. Geophysical Journal International, 103 680 (2), 363-376. 681 Avellán, D.R., Macías, J.L., Layer, P.W., Cisneros, G, Sánchez-Núñez, J.M., Gómez- 682 Vasconcelos, M.G., Pola, A., Sosa-Ceballos, G., García-Tenorio, F., Reyes-Agustín, G., 683 Osorio-Ocampo, S., García-Sánchez, L., Mendiola, I.F., Marti, J., López-Loera H., Benowitz, 684 J., 2018. Geology of the late Pliocene – Pleistocene Acoculco caldera complex, eastern Trans- 685 Mexican Volcanic Belt (México). Journal of Maps, 15(2), 8-18. 686 Bibby, H.M., Caldwell, T.G., Davey, F.J., Webb, T.H., 1995. Geophysical evidence on the 687 structure of the Taupo Volcanic Zone and its hydrothermal circulation. Journal of 688 Volcanology and Geothermal Research, 68 (1-3), 29-58. 689 690 Bursik, M., 2009. A general model for tectonic control of magmatism: Examples from Long Valley Caldera (USA) and El Chichón (México). Geofísica internacional, 48(1), 171-183. 691 Caballero, M., Lozano, S., Ortega, B., Urrutia, J., Macias, J.L., 1999. Environmental 692 characteristics of Lake Tecocomulco, northern basin of México, for the last 50,000 693 years. Journal of Paleolimnology, 22 (4), 399-411. 694 Calcagno, P., Evanno, G., Trumpy, E., Gutiérrez-Negrín, L.C., Macías, J.L., Carrasco-Núñez, G., 695 Liotta, D., the GEMex T3.1 team, 2018. Preliminary 3-D geological models of Los Humeros 696 and Acoculco geothermal fields (Mexico) – H2020 GEMex Project. Advances in Geosciences, 697 1, 1-13. 698 Campos-Enríquez, J.O., Alatriste-Vilchis, D.R., Huizar-Álvarez, R., Marines-Campos, R., 699 Alatorre-Zamora, M.A., 2003. Subsurface structure of the Tecocomulco sub-basin 700 (northeastern México basin), and its relationship to regional tectonics. Geofísica Internacional, 701 42 (1), 3-24. 702 Canet, C., Hernández-Cruz, B., Jiménez-Franco, A., Pi, T., Peláez, B., Villanueva-Estrada, R.E., 703 Salinas, S., 2015a. Combining ammonium mapping and short-wave infrared (SWIR) 704 reflectance spectroscopy to constrain a model of hydrothermal alteration for the Acoculco 705 geothermal zone, Eastern México. Geothermics, 53, 154-165. 29 706 Canet, C., Trillaud, F., Prol-Ledesma, R.M., González-Hernández, G., Peláez, B., Hernández- 707 Cruz, B., Sánchez-Córdova, M.M., 2015b. Thermal history of the Acoculco geothermal 708 system, eastern México: Insights from numerical modeling and radiocarbon dating. Journal of 709 Volcanology and Geothermal Research, 305, 56-62. 710 711 712 713 Castro-García, A., Córdoba, D., 1994. Estratigrafía del área volcánica de Tulancingo, Hidalgo, México. Revista de Investigación, Universidad Autónoma del Estado de Hidalgo, 1, 26-39. Costa, A., Martí, J., 2016. Stress field control during large caldera-forming eruptions. Frontiers in Earth Science, 4 (92), 1-13. 714 De Cserna, Z.D., De la Fuente-Duch, M., Palacios-Nieto, M., Triay, L., Mitre-Salazar, L. M., 715 Mota-Palomino, R., 1988. Estructura geológica, gravimetría, sismicidad y relaciones 716 neotectónicas regionales de la cuenca de México. Boletín del Instituto de Geología, 717 Universidad Nacional Autónoma de México, 104, 71. 718 719 De la Cruz-Martínez, V., Castillo-Hernández, D., 1986. Geología de la zona geotérmica de la caldera de Acoculco, Puebla. Geotermia, Revista Mexicana de Geoenergía, 2 (3), 245-254. 720 Di Napoli, R., Martorana, R., Orsi, G., Aiuppa, A., Camarda, M., De Gregorio, S., Candela E.G., 721 Luzio D., Messina, N., Pecoraino, G., Bitetto, M., de Vita, S., Valenza M., 2011. The structure 722 of a hydrothermal system from an integrated geochemical, geophysical, and geological 723 approach: The Ischia Island case study. Geochemistry, Geophysics, Geosystems, 12 (7), 1-25. 724 De Vicente, G., Muñoz, A., Giner, J.L., 1992. El uso del método de los diedros rectos: 725 implicaciones a partir del modelo de deslizamiento del análisis poblacional de fallas. Revista 726 Sociedad Geológica de España, 5, 3-4. 727 Ellis, R.G., De Wet, B., MacLeod, I.N., 2012. Inversion of magmatic data for remanent and 728 induced sources: Auxiliary Information in Geophysical Inversion. Presented at the ASEG 729 Inversion Forum, extended abstract. 730 731 García-Estrada, G.H., 2000. Gravimetría y magnetometría regional del proyecto Acoculco, Puebla. CFE-GPG internal report DEX-ACO19/2000, 57. 732 García-Palomo, A., Macías, J.L., Garduño, V.H., 2000. Miocene to Recent structural evolution 733 of the Nevado de Toluca volcano region, central México. Tectonophysics, 318 (1), 281-302. 734 García-Palomo, A., Macías, J.L., Tolson, G., Valdez, G., Mora, J.C., 2002. Volcanic stratigraphy 735 and geological evolution of the Apan región, east-central sector of the Trans-Mexica Volcanic 736 Belt. Geofísica Internacional, 41 (2), 133-150. 30 737 García-Palomo, A., Macías, J.L., Jiménez, A., Tolson, G., Mena, M., Sánchez-Núñez, J.M., 738 Arce, J.L, Layer, P.W., Santoyo, M.A., Lermo-Samaniego, J., 2018. NW-SE Pliocene- 739 Quaternary extension in the Apan-Acoculco region, eastern Trans-Mexican Volcanic Belt. 740 Journal of Volcanology and Geothermal Research, 349, 240-255. 741 García-Sánchez, L., Macías J.L., Arce, J.L., Sosa-Ceballos, G., Garduño-Monroy, V.H., 742 Saucedo, R., Avellán, D.R., Rangel, E., Layer, P.W. Lopez-Loera, H., Rocha, V.S., Cisneros, 743 G., Reyes-Agustín; G., Jiménez, A., 2016. Genesis and evolution of the Cerro Prieto Volcanic 744 Complex, Baja California, Mexico. Bulletin of Volcanology, 79 (44), 1-28. 745 García-Tovar, G.P., Martínez-Serrano, R.G., Solé, J., Correa-Tello, J.C., Nuñez-Castillo, E.Y., 746 Guilluo, H., Monroy-Rodríguez, E., 2015. Geología, geocronología y geoquímica del 747 vulcanismo Plio-Cuaternario del Campo Volcánico Apan-Tecocomulco, Faja Volcánica 748 Trans-Mexicana. Revista Mexicana de Ciencias Geológicas, 32 (1), 100-122. 749 Geyer, A., Martí, J., 2014. A short review of our current understanding of the development of 750 ring faults during collapse caldera formation. Frontiers in Earth Science, Volcanology, 2, 1-22. 751 Henry, C.D., Aranda-Gomez, J.J., 1992. The real southern Basin and Range: Mid-to late 752 Cenozoic extension in México. Geology, 20 (8), 701-704. 753 Hernández-Javier, I., 2008. Geología y geomorfología volcánica de la región de los yacimientos 754 de obsidiana en Otumba en el sector norte de la Sierra Nevada de México. Bacherlor thesis, 755 Universidad Nacional Autónoma de México. 756 Hiriart, G., Gutiérrez, N. L., Quijano, L.L., Ornelas, C.A., Espíndola, S., Hernández, I., 2011. 757 Evaluación de la energía geotérmica en México. Informe para el Banco Interamericano de 758 Desarrollo y la Comisión Reguladora de Energía, México. 759 Huízar-Álvarez, R., Campos-Enríquez, J.O., Lermo-Samaniego, J., Delgado-Rodríguez, O., 760 Huidobro-González, A., 1997. Geophysical and hydrogeological characterization of the sub- 761 basins of Apan and Tochac (México Basin). Geofísica Internacional, 36, 217-234. 762 Izquierdo, G., Aragón, A., Portugal, E., Arellano, V.M., De León, J., Álvarez, J., 2006. 763 Mineralogía de la zona mineralizada de sílice-epidota (ZMSE) del yacimiento geotérmico de 764 Cerro Prieto, B. C., México. Geotermia, Revista Mexicana de Geoenergia, 19, 2-12. 765 Lorenzo-Pulido, C., Armenta-Flores, M., Ramírez-Silva, G., 2010. Characterization of the 766 Acoculco Geothermal Zone as a HDR System. Geothermal Resources Council Transactions, 767 34, 369–372. 31 768 769 770 771 Johnson, C.A., Harrison, C.G.A., 1990. Neotectonics in central México. Physics of the Earth and Planetary Interiors, 64, 187-210. Layer, P.W., 2000. Argon-40/argon-39 age of the El’gygytgyn impactevent, Chukotka, Russia. Meteorite Planetary Science, 35, 591–599. 772 Layer, P.W., García-Palomo, A., Jones, D., Macías, J.L., Arce, J.L., Mora, J.C., 2009. El 773 Chichón volcanic complex, Chiapas, México: stages of evolution based on field mapping and 774 40Ar/39Ar geochronology. Geofisica Internacional, 48, 33–54. 775 Ledezma-Guerrero, O., 1987. Hoja Calpulalpan 14Q-h (3), con Resumen de la geología de la 776 hoja Calpulalpan, Estados de México, Puebla, Tlaxcala e Hidalgo. Quetzalapa region, southern 777 México. Revista Mexicana de Ciencias Geológicas, 15, 167-180. 778 Lermo, J., Antayhua, Y., Bernal, I., Venegas, S., Arredondo, J., 2009. Monitoreo sísmico en la 779 zona geotérmica de Acoculco, Puebla, México. Geotermia, Revista Mexicana de Geoenergía, 780 22 (1), 40-58. 781 782 Lipman, P.W., 2000. Calderas. In: Sigurdsson, H., Houghton, B.F., McNutt, S.R., Rymer, H., Stix, J. (Eds.), Encyclopedia of Volcanoes. Academic press, San Diego, 331-343. 783 López-Hernández, A., 2009. Evolución volcánica del complejo Tulancingo-Acoculco y su 784 sistema hidrotermal, Estados de hidalgo y Puebla, México. Ph.D. thesis, Universidad Nacional 785 Autónoma de México. 786 Lopez-Hernandez, A., Castillo-Hernandez, D., 1997. Exploratory drilling at Acoculco, Puebla, 787 México: a hydrothermal system with only nonthermal manifestations (No. CONF-971048). 788 Geothermal Resources Council, Davis, CA (United States). 789 López-Hernández, A., García-Estrada, G., Aguirre-Díaz, G., González-Partida, E., Palma- 790 Guzmán, H., Quijano-León, J., 2009. Hydrothermal activity in the Tulancingo-Acoculco 791 Caldera Complex, central México - Exploratory studies. Geothermics, 38, 279-293. 792 López-Hernández, A., Martínez, E.I., 1996. Evaluación volcanológica y estructural de la zona 793 geotérmica de Acoculco, Puebla, y su relación con la anomalía termal detectada en el pozo 794 EAC-1. CFE-GPG Internal Report OGL-AC-11/96. 795 Macías, J.L., Arce, J.L., García-Tenorio, F., Layer, P.W., Rueda, H., Reyes-Agustin, G., 796 Avellán, D., 2012. Geology and geochronology of Tlaloc, Telapón, Iztaccíhuatl, and 797 Popocatépetl volcanoes, Sierra Nevada, central México. Field Guides, 25, 163-193. 32 798 799 800 801 802 803 Martí, J., Geyer, A., Folch, A., Gottsmann, J., 2008. A review on collapse caldera modelling. Developments in Volcanology, 10, 233-283. Martí, J., Geyer, A., Aguirre-Diaz, G., 2013. Origin and evolution of the Deception island caldera (South Shetland Islands, Antarctica). Bulletin of Volcanology, 75 (6), 732. McDougall I., Harrison, T.M., 1999. Geochronology and thermochronology by the 40Ar/39Ar method, 2nd edition. Oxford University Press, New York. 804 Molina, F., Martí, J., Aguirre, G., Vega, E., Chavarría, L., 2014. Stratigraphy and structure of the 805 Cañas Dulces caldera (Costa Rica). Geological Society of America Bulletin, 126 (11-12), 806 1465-1480. 807 808 Mooser, F., Ramírez, M.T., 1987. Faja Volcánica Transmexicana: Morfoestructura, tectónica y vulcanotectónica. Boletín de la Sociedad Geológica Mexicana, 48 (2), 69-73. 809 Pardo, M., Suárez, G., 1995. Shape of the subducted Rivera and Cocos plates in southern 810 México: Seismic and tectonic implications. Journal Geophysical Research, 100 (12), 357-12. 811 Peiffer, L., Bernard-Romero, R., Mazot, A., Taran, Y.A., Guevara, M., Santoyo, E., 2014. Fluid 812 geochemistry and soil gas fluxes (CO 2–CH 4–H 2 S) at a promissory Hot Dry Rock 813 Geothermal System: The Acoculco caldera, México. Journal of Volcanology and Geothermal 814 Research, 284, 122-137. 815 Pérez‐Campos, X., Kim, Y., Husker, A., Davis, P.M., Clayton, R.W., Iglesias, A., Pacheco, J., 816 Singh, S., Constantin, M.V., & Gurnis, M., 2008. Horizontal subduction and truncation of the 817 Cocos Plate beneath central Mexico. Geophysical Research Letters, 35(18). 818 Polak, B.G., Prasolov, E.M., Kononov, V.I., Verkhocskiy, A.B., González, A., Templos, L. A., 819 Mañón, A., 1982. Isotopic composition and concentration of inert gases in Mexican 820 hydrothermal systems (genetic and applied aspects). Geofísica Internacional, 21 (3), 193-227. 821 Rocha, S., Jiménez, E., Palma, H., 2006. Propuesta para dos pozos exploratorios en el proyecto 822 geotérmico de Acoculco, Puebla. Comisión Federal de Electricidad, internal report OGL- 823 ACO-03/06, unpublished. 824 Renne P.R., Mundil, R., Balco, G., Min, K., Ludwig, K.R., 2010. Joint determination of 40K 825 decay constants and 40Ar*/40K for the Fish Canyon sanidine standard, and improved 826 accuracy for 40Ar/39Ar geochronology. Geochimica et Cosmochimica Acta, 74, 5349–5367 33 827 Rueda, H., Macías, J.L., Arce, J.L., Gardner, J.E., Layer, P.W., 2013. The~ 31ka rhyolitic 828 Plinian to sub-Plinian eruption of Tlaloc Volcano, Sierra Nevada, central México. Journal of 829 Volcanology and Geothermal Research, 252, 73-91. 830 Ryan, W.B., Carbotte, S.M., Coplan, J.O., O'Hara, S., Melkonian, A., Arko, R., Weissel, R.A., 831 Ferrini, V., Goodwillie, A., Nitsche, F., Bonczkowski, J., 2009. Global multi‐resolution 832 topography synthesis. Geochemistry, Geophysics, Geosystems, 10 (3). 833 834 Saxby, J., Gottsmann, J., Cashman, K., & Gutiérrez, E., 2016. Magma storage in a strike-slip caldera. Nature communications, 7, 12295. 835 Siebe, C., Macias, J.L., Abrams, M., Rodriguez, S., Castro, R., Delgado, H., 1995. Quaternary 836 explosive volcanism and pyroclastic deposits in east central México: implications for future 837 hazards. In: Chacko, J., Whitney A. (Eds.), Guidebook of geological excursions, in 838 conjunction with the Annual Meeting of the Geological Society of America. New Orleans, 839 Louisiana, 1-48. 840 Sosa-Ceballos, G., Macías, J.L., Avellán, D.R., Salazar-Hermenegildo, N., Boijseauneau-López, 841 M.E., Pérez-Orozco, J.D., 2018. The Acoculco Caldera Complex magmas: Genesis, evolution 842 and relation with the Acoculco geothermal system. Journal of Volcanology and Geothermal 843 Research, 358, 288-306. 844 Suter, M., 1984. Cordilleran deformation along the eastern edge of the Valles–San Luis Potosí 845 carbonate platform, Sierra Madre Oriental fold-thrust belt, east-central Mexico. Geological 846 Society of America Bulletin, 95(12), 1387-1397.Suter, M., Carrillo-Martínez, M., Quintero- 847 Legorreta, O., 1996. Macroseismic study of shallow earthquakes in the central and eastern 848 parts of the trans-Mexican volcanic belt, México. Bulletin Seismological Society of 849 America, 86 (6), 1952-1963. 850 851 852 853 Tello-Hinojosa, E., 1986. Geoquímica de la zona geotérmica de la caldera de Acoculco, Puebla. CFE-GPG internal report 34/86, 15. Urrutia-Fucugauchi, J., Flores-Ruiz, J., 1996. Bouger gravity anomalies and regional crustal structure in central Mexico. International Geology Review, 38, 176–194. 854 Viggiano-Guerra, J.C., Flores-Armenta, M., Ramírez-Silva, G.R., 2011. Evolución del sistema 855 geotérmico de Acoculco, Pue., México: un estudio con base en estudios petrográficos del pozo 856 EAC-2 y en otras consideraciones. Geotermia, Revista Mexicana de Geoenergia, 24 (1), 14- 857 24. 34 858 Zúñiga, F.R., Pacheco, J.F., Guzmán-Speziale, M., Aguirre-Dıaz, G.J., Espındola, V.H., Nava, 859 E., 2003. The Sanfandila earthquake sequence of 1998, Querétaro, México: activation of an 860 undocumented 861 Belt. Tectonophysics, 361 (3), 229-238. fault in the northern edge of central Trans-Mexican Volcanic 862 863 FIGURE CAPTIONS 864 Figure 1. A) Regional tectonic configuration of Central México showing the active subduction of 865 the Rivera and Cocos Plates beneath the North American Plate at the Middle-American Trench. 866 The dotted black line depicts the boundary of the Trans-Mexican Volcanic Belt (TMVB) and the 867 location of the Acoculco Caldera Complex and other volcanoes (LP, La Primavera; C, Fuego de 868 Colima; T, Tancítaro; Am, Amealco; Hc, Huichapan; To, Nevado de Toluca; P, Popocatepetl; M, 869 Malinche; O, Pico de Orizaba; and Hm, Humeros). The shaded relief map and bathymetric data 870 was acquired and modified from Ryan et al. (2009). 871 Figure 2. Simplified geological map of the Acoculco Caldera Complex after Avellán et al. 872 (2018), Sosa-Ceballos et al. (2018), and Calcagno et al. (2018). The map contains the distribution 873 of 30 stratigraphic units and a composite stratigraphic column supported by 24 new 874 ages and two 14C dates (Fig. 3). This map shows towns, main roads, the location of stratigraphic 875 logs, samples with age determinations, faults, the zone of intense hydrothermal alteration, and the 876 two CFE exploratory wells. The DEM was obtained from the INEGI topographic data with a 877 horizontal resolution of 15-m. The geologic profile at the bottom correlates the surface units of 878 this map with the CFE stratigraphic logs of the EAC-1 and EAC-2 wells with the location of 879 shallow intrusions. 880 Figure 3. Detailed stratigraphic column of the units and eruptive stages of the Acoculco Caldera 881 Complex composed of 30 units marked in the geological map of figure 2. To the right appear the 882 plateau age of units, the plateau age sorted by age, and their estimated total volume. 35 40Ar/39Ar 40Ar/39Ar 883 Figure 4. 884 Caldera Complex (see Table 1). Plateau and Isochron ages, where calculated, are reported at the 885 1-sigma level. For Ac90, the 6 single-step fusion results are shows as a single probability density 886 histogram (see Table 1). 887 Figure 5. Photos of selected outcrops of the Acoculco andesitic ignimbrite (Aai) with its different 888 lithological facies. A, Aspect of the welded massive structure of the ignimbrite with pumice 889 fragments embedded in a light-yellowish ash matrix. Hand lens for scale. B, Lacustrine deposits 890 that partly covered the Aai with parallel plane stratification. They are made of interbedded 891 massive fine sandy beds with laminar-stratified clayed beds. Notice that the succession is gently 892 tilted. C, Co-ignimbrite lag breccia (mlBr) of the Acoculco ignimbrite transected by E-W 893 structures and NW-SE faults. Notice that the lava lithics are supported by a light-brown to 894 yellowish ashy matrix with pumice. D, Co-ignimbrite lag breccia (mlBr) at the bottom of the 895 outcrop in contact with the Acoculco ignimbrite. The color change marks a difference in the 896 texture and lithological components of the deposits. 897 Figure 6. 898 Caldera Complex (see Table 1). Plateau and Isochron ages, where calculated, are reported at the 899 1-sigma level. 900 Figure 7. A, Regional structural map modified from García-Palomo et al. (2018). Dark gray areas 901 correspond to horsts and light gray areas correspond to grabens. Dike, scoria cones and 902 slickensides symbols show extension direction. B, Structural map of the Acoculco Caldera 903 Complex showing the main structural features, field sites (outcrops) and the caldera rim (this 904 study). C, The fault planes were plotted on stereographic projections in the lower hemisphere. 905 The trend-frequency polar diagram (rose diagram) was obtained with Rozeta 2.0, and the fault 906 planes and poles were plotted with Stereonet 9. D, A slip model showing an orthorhombic 40Ar/39Ar age spectra and inverse isochron diagrams of dated rocks of the Acoculco age spectra and inverse isochron diagrams of dated rocks of the Acoculco 36 907 symmetry of NW- and NE- oriented faults moving under the same strain field. Black zones 908 correspond to 100% compatible compression, and white zones 100% tension areas. 909 Figure 8. A, Image of the Magnetic Field Reduced to the Pole (MFRP) overlaid on a DEM of the 910 ACC region. The black dots represent the CFE wells. B, A 3D inversion model of the magnetic 911 susceptibility produced with the Geosoft´s VOXI software by using the magnetic vector inversion 912 method (Ellis et al., 2012). In this image four anomalies related to intrusive bodies occur at 913 depths between 1,000 and >2,500 m below the surface. 914 Figure 9. Eruptive stages during the formation of the Acoculco Caldera Complex described as 915 pre-caldera (A), syn-caldera (B), early post-caldera (C) late post-caldera (D), and still late post- 916 caldera and extra-caldera (E). After the caldera collapse two episodes of caldera reactivation 917 generated the Encimadas (Eri), and Tecoloquillo (Tr1) rhyolitic ignimbrites. Outside of the 918 caldera rim, the products of the Acoculco caldera are interbedded with deposits of the Apan- 919 Tezontepec volcanic field (extra caldera volcanism). 920 LIST OF TABLES 921 Table 1. List of dated samples from rock units of the Acoculco Caldera Complex. See text for 922 map units. All ages (in ka) are reported at the 1-sigma level. For most samples, the 40Ar/39Ar 923 plateau age is the interpreted age (shown in bold) of the unit. Sample Aco901 showed evidence 924 of excess argon, so the isochron age is the interpreted age. The age of sample Ac90 is a weighted 925 mean age of 6 single-step fusions, and for sample Ac94a, no plateau or isochron could be 926 calculated. MSWD: Mean Square Weighted Deviation. See text for analytical methods and 927 standards used. List of rock units of the Acoculco Caldera Complex dated in this study and their 928 40Ar/39Ar 929 Table 2. Radiocarbon date of paleosol sample performed during study. The location is in UTM 930 coordinates. Analyses performed at the International Chemical Analysis Inc. Florida, USA. ages. 37 931 38 Table 1 continued Table 1. Sample ACC stage Unit Location North East Material Integrated Age (ka) Plateau Age (ka) Plateau information Isochron Age (ka) Isochron information Fractions=7 Init. 40Ar/36Ar=291.2 ±3.2 MSWD=1.83 Fractions=10 of 14, 2 runs Init. 40Ar/36Ar=303.4 ±0.8 MSWD=0.46 Fractions=6 Init. 40Ar/36Ar=297.2 ±12.1 MSWD=0.64 Ac42 Late postcaldera Plc 2199245 593432 Whole rock 28 ± 12 71 ± 17 Fractions=7 79.0% of 39Ar rel. MSWD=2.23 121 ± 32 Aco901 Late postcaldera aplitic dike 2203029 589693 Whole rock - - No plateau 183 ± 36 Ac11 Late postcaldera Tr1 2192344 586149 Plagioclase 614 ± 82 611 ± 72 Fractions=5 99.2% of 39Ar rel. MSWD=0.69 633 ± 72 Ac11 Late postcaldera Tr1 2192344 586149 Pumice 756 ± 10 762 ± 9 Fractions=8 %99.4 of 39Ar rel. MSWD=0.71 755 ± 13 Fractions=8 Init. 40Ar/36Ar=299.4 ±9.0 MSWD=0.77 Ac104 Late postcaldera Arcd 2194852 591775 Whole rock 921 ± 38 998 ± 36 Fractions=6 92.4% of 39Ar rel. MSWD=0.49 1,171 ± 81 Fractions=7 Init. 40Ar/36Ar=287.5 ±3.5 MSWD=0.28 Ac89 Late postcaldera Mrcd 2208434 572356 Whole rock 1,363 ± 54 1,066 ± 42 Fractions=5 78.7% of 39Ar rel. MSWD=2.23 1,036 ± 43 Fractions=8 Init. 40Ar/36Ar=304.4 ±2.3 MSWD=2.29 Ac37 Late postcaldera Msf 2199284 575914 Pumice 2,850 ± 197 1,084 ± 22 Fractions=4 87.1% of 39Ar rel. MSWD=2.45 1,136 ± 49 Fractions=4 Init. 40Ar/36Ar=290.3 ±6.2 MSWD=2.38 Ac82 Late postcaldera Arcd 2190307 595629 Obsidian 1,139 ± 10 1,145 ± 14 Fractions=8 99.8% of 39Ar rel. MSWD=2.29 1,130 ± 9 Fractions=8 Init. 40Ar/36Ar=300.5 ±5.0 MSWD=1.69 Ac100 Late postcaldera Eri 2204497 596510 Whole rock 1,240 ± 12 1,278 ± 14 Fractions=5 94.4% of 39Ar rel. MSWD=1.83 - No isochron Ac72 Late postcaldera Ard 2199502 578889 Whole rock 1,190 ± 54 1,283 ± 88 Fractions=4 78.9% of 39Ar rel. MSWD=2.08 1,309 ± 526 Fractions=4 Init. 40Ar/36Ar=295.9 ±17.6 MSWD=2.92 Ac92 Late postcaldera Trcd 2213879 583057 Whole rock 1,329 ± 12 1,360 ± 15 Fractions=4 76.0% of 39Ar rel. MSWD=1.83 1,392 ± 27 Fractions=6 Init. 40Ar/36Ar=285.3 ±9.2 MSWD=3.04 Ac94B Late postcaldera Crcd 2216042 585241 Whole rock 1,339 ± 8 1,394 ± 8 Fractions=5 71.6% of 39Ar rel. MSWD=1.22 1,410 ± 12 Fractions=5 Init. 40Ar/36Ar=289.3 ±4.6 MSWD=0.97 Ac107 Late postcaldera Amrc 2213875 590387 Whole rock 1,413 ± 27 1,438± 24 Fractions=6 95.9% of 39Ar rel. MSWD=0.20 1,462 ± 59 Fractions=6 Init. 40Ar/36Ar=273.2 ±58.6 MSWD=0.21 Table 1 continued Ac98B Late postcaldera Cbd 2202522 592368 Whole rock 1,562 ± 42 1,600 ± 35 Fractions=6 97.2% of 39Ar rel. MSWD=0.57 1,608 ± 32 Fractions=7 Init. 40Ar/36Ar=294.1 ±2.3 MSWD=0.63 Ac76 Late postcaldera Tal 2201609 587830 Whole rock 1,717 ± 40 1,708 ± 54 Fractions=4 90.5% of 39Ar rel. MSWD=2.64 1,738 ± 26 Fractions=8 Init. 40Ar/36Ar=294.9 ±2.8 MSWD=1.55 Ac113 Late postcaldera Alrd 2196999 584295 Whole rock 1,864 ± 38 1,870 ± 36 Fractions=7 99.6% of 39Ar rel. MSWD=0.71 1,854 ± 32 Fractions=7 Init. 40Ar/36Ar=285.1 ±80.2 MSWD=0.78 Ac69B Syncaldera Aai 2200645 584104 Pumice 1,991 ± 33 2,041 ± 38 Fractions=5 96.2% of 39Ar rel. MSWD=1.59 2,018 ± 56 Fractions=5 Init. 40Ar/36Ar=299.3 ±8.9 MSWD=1.98 Ac44 Early postcaldera Atad 2195570 592761 Whole rock 2,183 ±19 2,179 ± 26 Fractions=4 82.3% of 39Ar rel. MSWD=2.33 2,207 ± 31 Fractions=8 Init. 40Ar/36Ar=288.9 ±12.8 MSWD=4.76 Ac69A Syncaldera Aai 2200645 584104 Pumice 2,279 ± 86 2,185 ± 65 Fractions=3 87.7% of 39Ar rel. MSWD=0.71 2,113 ± 100 Fractions=8 Init. 40Ar/36Ar=308.2 ±9.4 MSWD=0.54 Ac80 Early postcaldera Mtal 2198103 586869 Whole rock 2,113 ± 28 2,199 ± 24 Fractions=4 87.2% of 39Ar rel. MSWD=0.97 2,316 ± 70 Fractions=4 Init. 40Ar/36Ar=255.6 ±26.6 MSWD=0.14 Ac103 Early postcaldera Vtal 2205459 593663 Whole rock 2,288 ± 53 2,323 ± 48 Fractions=7 99.1% of 39Ar rel. MSWD=0.35 2,321 ± 47 Ac90 Early postcaldera Srl 2208919 574826 - 2,553 ± 110 - - - Ac94A TMVB Pald 2216042 585241 Whole rock 3,620 ± 22 - No plateau - No isochron Ac15 TMVB Pdhd 2190111 579819 Whole rock 10.74 ± 0.06 Ma 10.98 ± 0.07 Ma Fractions=4 56.4% of 39Ar rel. MSWD =2.16 - No isochron Fractions=7 Init. 40Ar/36Ar=295.9 ±4.8 MSWD=0.42 Single shot fusions on 6 chips, MSWD=1.15 Probability=0.33 Table 2 Unit Clc Outcrop Description (sample) Ac1 Paleosol below Clc Location North East 2197465 594866 Lab. Code Pretreatment Conventional age Calibrated Age 18OS/0373 AO 16,710 ± 50 BP Cal 18410 - 18020 BC

RELATED PAPERS

RELATED TOPICS

Log In

Eruptive chronology of the Acoculco caldera complex – A resurgent caldera in the eastern Trans-Mexican Volcanic Belt (México)

Eruptive chronology of the Acoculco caldera complex – A resurgent caldera in the eastern Trans-Mexican Volcanic Belt (México)

Related Papers

RELATED PAPERS

RELATED TOPICS