Seismicity and Volcanic Activity of Ceboruco Volcano (Mexico): Insights from Historical and Recent Studies

1. Introduction

Ceboruco Volcano (21° 07′ 30″ N and 104° 30′ 29″ W; 2,280 m asl) is located in central-western México, in the state of Nayarit. It is situated at the northwestern edge of the Tepic-Zacoalco Rift (TZR), at the northwest edge of the Trans-Mexican Volcanic Belt (TMVB), close to the boundary with the Sierra Madre Occidental (SMO) (Figure 1a) [1, 2]. Two concentric horseshoe-shaped calderas form Ceboruco Volcano, with one larger outer caldera and a second, smaller inner caldera [3–5]. Within the inner caldera are several domes, and andesitic to basaltic lava flows and cinder cones comprise the edifice (Figure 1b), which has a generally NW–SE trend [4, 5]. Adjacent to the Ceboruco Volcano are several small monogenetic volcanic formations, which include San Pedro Pochotero, El Comal, El Pinchancha, El Molcajete, and El Pedregal. Located 42 kilometers to the northwest is Sangangüey Volcano, another site recognized for its potential volcanic activity within the region (see Figure 1a). The eruptive history of Ceboruco has been classified into four distinct stages [6–8]. The principal edifice may have attained an elevation of approximately 2,700 meters above sea level and is primarily characterized by andesitic lava flows, which exhibit an absence of pyroclastic deposits within the interstratified layers, corresponding to the initial stage of its development (see Figure 1b). This initial volcanic cycle concluded roughly 1,000 years ago with a Plinian eruption that produced rhyodacitic pumice (the Jala eruption), resulting in the formation of the first outer caldera (see Figure 1c).

The second stage is the formation of the Dos Equis dacitic dome within the first caldera, which reached a size of 1.78 km in diameter and 280 m above the caldera floor, characterizing the second stage (Figure 1b) [7]. It was proposed that the second, internal caldera was formed by the collapse of the Dos Equis dome, but there is no agreement among researchers on the age or mechanism of this collapse [6, 9–11].

The inception of the third eruptive cycle was marked by the emplacement of an andesitic dome within the intracaldera region. This phase involved effusive andesitic events, which are defined geochemically by a more evolved (acidic) composition and enriched concentrations of titanium ($TiO_2$) and potassium ($K_2O$) oxides [8]. Subsequently, the fourth stage encompasses recent volcanism originating from the inner caldera, characterized by the extrusion of two dacitic domes and the emplacement of the historic 1870 lava flow [7, 10, 12, 13].

Since the Jala eruption, which is characterized by a Volcanic Explosivity Index of 6 in the year [1,065 ± 55 years] [13], a minimum of seven notable eruptive events have transpired [5, 14]. This frequency indicates an average interval of 126 years between eruptions (see Figure 1c), with the most recent event occurring between 1870 and 1872 (see Figure 2). Given the nature and recurrence of Ceboruco’s eruptive episodes, its associated volcanic hazards rank among the highest in Mexico, highlighting the urgent requirement for ongoing monitoring of volcanic activity. Furthermore, the continuous growth of the population and economic activities in the vicinity of the volcano necessitates a comprehensive assessment of hazards and risks, akin to evaluations conducted in Chile after the eruption of the Chaitén volcano. In this study, we present findings from a systematic seismic survey of Ceboruco Volcano aimed at characterizing its seismicity and investigating its relationship with the primary regional structural features.

2. Tectonic framework

The TMVB originated during the Middle Miocene and continues to the present [15–17]. The volcanism along the TMVB is due to the subduction of the Rivera and Cocos plates beneath North America. The oblique orientation of the TMVB relative to the Middle America Trench gives rise to intricate plate interactions that generate a complex magmatic history [16]. The volcanic rocks have calc-alkaline to alkaline affinities, ranging from mafic to acidic, and the volcanic edifices range from stratovolcanoes to monogenetic volcanoes and domes [15, 16, 18].

Deformation in the study area is, to date, associated with the TZR evolution. The origin of the TZR structure is linked to the opening of the Gulf of California during the Late Miocene [19]. This rift is NW–SE-trending and composed of segments of normal faulting, along some of which volcanism has occurred [19, 20]. The stress regime is extensional, with σ3 oriented EW to ENE–WSW, with a lateral component [21–23]. Whether the lateral component is right-lateral (references) or left-lateral [16], consistent with the Jalisco Block motion of ~2 mm/year to the southwest [24], local evidence of accommodation zones has been observed [16].

Structural lineaments with a northwest–southeast (NW–SE) trend have served as preferential pathways for magmatic propagation and emplacement at diverse spatial scales, manifesting as monogenetic fields, dacitic domes, and composite stratovolcanoes [7, 14, 19]. Magmatism within the TZR exhibits temporal continuity extending from the Pleistocene to the Holocene, typified by the Ceboruco volcanic complex and its satellite vents. Stratigraphic evidence from the early evolutionary stages indicates a strong tectonic control, characterized by the alignment of multiple eruptive episodes parallel to the NW–SE structural fabric of the TZR [7, 14, 25].

3. Structural features

Prior investigations of the Ceboruco Volcano have predominantly focused on lithostratigraphy and evolutionary chronology. While the regional structural architecture and tectonic implications of the TZR are well documented, detailed descriptions of local structural constraints remain limited. Ferrari et al. [19] identify WNW-trending alignments of scoria cones and domes defining the boundaries of the San Pedro-Ceboruco graben; these features, colloquially designated as the northern and southern volcanic chains, exhibit clear fault-controlled emplacement. Although the NW–SE structural grain is dominant, the region possesses a complex polyphase tectonic history dating to the Late Miocene, characterized by a heterogeneous array of fault systems-oriented N–S, NNW–SSE, E–W, and ENE–WSW [17, 21, 23, 26]. While a comprehensive genetic analysis of each fault system exceeds the scope of this study, we correlate the most recent faulting episodes with seismicity relevant to contemporary unrest at Ceboruco. Local structural analysis (Figure 3) reveals two prevailing trends: Late Miocene NW–SE faults [23] are crosscut by Plio-Pleistocene ENE–WSW structures [16]. The intersection of these distinct fault populations likely generated a zone of structural weakness or extension, facilitating magmatic ascent and the subsequent construction of the Ceboruco edifice. The influence of structural inheritance and crosscutting fault networks on volcanogenesis is well-documented in analogous rift settings (e.g., [27, 28]). Preexisting crustal discontinuities are critical determinants for magma reservoir emplacement and the morphology of resulting volcanic centers. Consequently, the tectonic configuration and deformation style characterizing the TZR provide a favorable environment for the magmatic propagation observed throughout the evolution of Ceboruco [7, 14, 19, 25].

4. Local geology

The lithostratigraphic framework of Ceboruco Volcano has been characterized with heterogeneous levels of resolution, resulting in the proliferation of informal nomenclature. While initial descriptions relied primarily on caldera evolution stages [7], subsequent revisions have redefined the sequence through an integrated analysis of contact relationships, geochronology, and geochemistry [14, 19, 29]. Consequently, the original classification, comprising 19 informal lithodemic units [7], has been consolidated into a contemporary consensus of eight distinct units [13, 18]. This refined division correlates lithological variations with specific geochemical signatures and shifts in the regional tectonic stress field. Figure 3 illustrates the stratigraphic column [14, 19], which spans the Pliocene–Holocene interval. The basal units consist of the Jala rhyolitic ignimbrites, dated between 4.9 and 4.1 Ma [19]. Stratigraphically overlying the Jala sequence is the Ixtlán and Buenavista mafic-to-intermediate complex (3.8 Ma), exposed along the southern periphery of the edifice. The genesis of the Ceboruco edifice proper commenced circa 50 ka, marked by the emplacement of monogenetic cones and domes within the northern volcanic chain; the NW–SE alignment of these features suggests strong structural control by regional fault systems. The incipient construction phase of the stratovolcano is represented by the ~45 ka precaldera lava series, which constitutes the bulk of the volcanic edifice. A significant explosive episode occurred ~1 ka, depositing the Marquezado tuff through the dome field on the southern flank. The terminal evolutionary stage is defined by the postcaldera lavas, culminating in the historical 1870 effusive eruption on the southwest flank (Figures 1b and 3).

5. Seismicity background

The first local studies of seismicity at Ceboruco Volcano were conducted in 1993–1994, following the installation of a temporary network comprising two MQ Spregnether analog stations and two Lennartz Mars88 digital stations equipped with LE-3D sensors. Between 1996 and 1998, the BloJal Project [30] was carried out to study seismicity in the Jalisco Block using five Lennartz Mars88 stations with LE-3D sensors in various network configurations over different periods. As a result of the BloJal Project [31], seismic activity at Ceboruco Volcano was identified.

In 2002, a collaborative initiative between the Research Group CA-UDG-276 (SisVOc) of the Universidad de Guadalajara and the Civil Defense Agency of the State of Nayarit established a semi-permanent, autonomous seismic monitoring station (designated CEBN) on the southern flank of Ceboruco Volcano. The instrumentation consisted of a Lennartz MARSLite data acquisition system coupled with a Lennartz LE-3D/1s (1 Hz) seismometer. Data were digitized at 125 Hz, with temporal synchronization provided by GPS locking at three-hour intervals. The primary objective of this deployment was to characterize the baseline seismicity of the volcanic edifice and to develop a classification framework for local seismic events. Operational continuity was frequently compromised by intermittent power supply and vandalism, resulting in the station’s permanent decommissioning in 2009.

Using the best data available from the CEBN station for the period 2003–2008 [32], researchers studied the seismic events at Ceboruco Volcano and identified volcano-tectonic (VT), low-frequency (LF), and hybrid earthquakes according to [32]. These authors calculated seismicity rates of 0.43 earthquakes per day (13 per month) for VT earthquakes and 0.31 earthquakes per day (9 per month) for LF earthquakes. They mainly analyzed VT earthquakes, using location techniques for a single three-component station (azimuth and amplitude of the first arrival). From these angles and the S–P times, approximate hypocentral locations were obtained, assuming typical volcano seismic velocities of Vp = 3.00 km/s and Vs = 1.73 km/s [32]. They found that most of the epicenters were distributed within a radius of 6 km around the station (2.5 km to the SW of the summit) (Figure 4).

Using the [32] database, [3] studied the LF or b-type earthquakes of the Ceboruco Volcano. These authors reported that the earthquakes had a wide range of forms, durations, and spectral contents; however, they were subclassified into four groups or families based on similar characteristics in the time and frequency domains: Short Duration, Extended Coda, Bobbin, and Modulated Amplitude (Figure 5) [3]. They located the low-frequency earthquakes described by [32] and found that most of the epicenters were randomly distributed within a radius of 9 km around the station (Figure 6). The complete set of seismograms, spectra, spectrograms, and particle-motion determinations is available in [3].

Between March 2012 and July 2014, a seismic monitoring campaign was conducted at Ceboruco Volcano utilizing a temporary network of four portable stations (network code: RCEB). Waveform analysis facilitated the discrimination of three distinct classes of volcanic seismicity (Figures 7–9). From a cumulative dataset of 489 recorded events, constrained hypocentral locations were derived for a subset of 33 earthquakes (Figure 10) based on the criterion of clear P- and S-phase arrivals recorded at a minimum of three stations. The P- and S-phase arrival times were obtained using particle motion. As in previous studies, almost all earthquakes occur within a radius of approximately 9 km of station CEBN. Most hypocenters are shallow, with depths between 0 and 10 km [33].

The spatial distribution of epicenter clusters proximal to the volcanic edifice delineates three discrete structural lineaments with a preferential ENE–WSW orientation (Figure 10). This structural geometry is orthogonal to the TZR axis and conforms to the characteristic patterns of the most recent neotectonic deformation phase. Hypocentral solutions indicate shallow focal depths confined to the upper 10 km of the crust, suggesting a seismogenic source driven by local tectonic stress accumulation, magmatic recharge dynamics, or a coupled tectonic-magmatic mechanism.

At the end of 2016, a dense temporary seismic network was deployed [34], spanning an area of 16 km x 16 km around Ceboruco. The 25 seismic stations operated until July 2017 as part of the P-24 project of the CeMIEGeo Consortium (Centro Mexicano de Innovación en Energía Geotérmica), whose goal was to study the local seismicity and the volcanic edifice structure (Figure 11).

A database was generated using Antelope software. Data processing and event classification were conducted through a multistage workflow:

• Automated detection: Initial event identification was performed using the Antelope software suite, utilizing a standard short-term average/long-term average energy-triggering algorithm to screen continuous data streams.

• Phase picking: Automated triggers underwent manual analyst review to refine arrival times. P-wave onsets were identified on vertical-component seismograms, while S-wave arrivals were picked on the horizontal components to ensure phase accuracy.

• Hypocentral determination: Earthquake locations were computed within the Antelope environment, constrained by the IASP91 one-dimensional reference velocity model.

About 56 earthquakes were selected, like the one shown in Figure 12, within the region (21.0°, 21.2°) N and (−104.65°, −104.40°) W.

Following the Ceboruco Volcano seismicity classification schemes by [3, 31, 32], we classified the events in our catalog into hybrid and low-frequency types. Low-frequency (LF) events, in turn, have been subclassified into Type I or Short Duration events (Figure 13a), Type II or Extended (Figure 13b), and Type IV or Modulated Amplitude earthquakes (Figure 14a), according to the classification made by [3]. Hybrid events (Figure 14b) are characterized by high-frequency onsets and low-frequency codas with spectral contents of approximately 2 to 10 Hz.

Visual inspection of the located event catalog revealed a high degree of waveform similarity, suggesting a repetitive source mechanism. To systematically quantify this similarity and identify repeating events, we employed the cross-correlation detector module within the ObsPy seismological library. High-fidelity master templates were selected from stations CB02, CB13, and CB19 based on their signal-to-noise ratios (Figure 15). The continuous data archive was scanned using a matched-filter technique governed by the following processing parameters:

• Spectral filtering: Application of a 1–15 Hz band-pass filter to enhance the signal of interest.

• Detection threshold: A minimum crosscorrelation coefficient (CC) of > 0.75 was required to declare the detection.

• Temporal constraint: A minimum inter-event separation time of 10 seconds was enforced to mitigate redundant triggering or duplicate detections.

Application of the crosscorrelation algorithm yielded 25 additional events, expanding the cumulative catalog to 81 seismic entries. Spatiotemporal and hypocentral distributions are illustrated in Figure 16. Depth frequency analysis reveals a predominant clustering between 6 and 8 km, accounting for 48 events (59.3%). Shallower seismicity (<6 km) comprises 23 events (28.4%), while events exceeding 8 km depth represent the remaining 12.3%. Notably, only a single event was located at a depth greater than 10 km. Cross-sectional profiles (Figure 16) depict spatiotemporal clustering, with multiple events occurring synchronously within narrow depth ranges, consistent with the phenomenological definition of a seismic swarm. Local magnitude ($M_L$) determination was feasible for 64 events (79% of the dataset). The majority fall within the range of 0.5 ≤ M_L ≤ 1.5.

A subset of 14 events (17.2%) exceeded M_L = 1.5, with only four events surpassing M_L > 2.0. The maximum magnitude event (M_L = 2.6) was recorded on April 27, 2017, at 17:41 local time (Figure 12). Analysis of hypocentral migration patterns (Figure 16a) indicates a temporal progression of seismicity along a NW–SE trajectory (Figure 16e), accompanied by an upward migration of focal depths from 8 km to 4 km (Figure 16c). Relocation metrics indicate high solution quality: 93% of events (n = 75) exhibit horizontal errors (ERH) <2.0 km, while 73% (n = 59) maintain vertical errors (ERZ) <2.0 km. Regarding root mean square (RMS) travel-time residuals, 65% of solutions fall within the 0.40–0.49 s interval, with 32.5% achieving an RMS <0.4 s.

A recent study using 3D inversion modeling of magnetotelluric data [35] proposes the existence of a conductive anomaly, C2 [2 km depth], beneath the Ceboruco Volcano summit, with resistivity values of 2–5 Ωm. This interpretation is intriguing and could correspond to a shallow ancient magma chamber enveloped by high-temperature fluids interacting with the host-rocks and hydrothermal system. The predominant spatial distribution of seismicity identified in this study clusters along the southwestern margin of the C2 anomaly (Figure 16b, c), a finding that corroborates prior interpretative models. However, the occurrence of events exhibiting reverse faulting mechanisms, coinciding with a distinct vertical migration of hypocenters from depth toward the surface, suggests a local compressional regime driven by ascending magmatic or hydrothermal fluids. These kinematic and spatiotemporal patterns may indicate a potential reactivation of the volcanic system.

6. Conclusion(s)

The integration of these analyses has facilitated the precise delineation of seismogenic zones associated with the Ceboruco volcanic complex, confirming that local seismicity is governed by the dynamic interplay between the magmatic system and the ambient tectonic stress field. A defining characteristic of the observed activity is its tendency to manifest as seismic swarms. This study constitutes the first high-resolution characterization of the microseismic regime in the volcano’s proximal field. The catalog reveals distinct spatiotemporal migration patterns, specifically a lateral progression from NW to SE, accompanied by a vertical shallowing of hypocenters. Furthermore, these data elucidate the role of local structural controls, identifying both active faults consistent with the TZR geometry and the fluid-driven reactivation of preexisting basement structures that are oblique to the contemporary deformation field.

Similar advances in seismic monitoring have been made in other volcanic regions worldwide. In Japan, for instance, Mount Fuji and Sakurajima are continuously monitored by extensive seismic and geodetic networks, which provide real-time data vital for assessing hazards in densely populated regions. Similarly, Mount Etna and Vesuvius in Italy are equipped with extensive arrays of seismic stations operated by the Istituto Nazionale di Geofisica e Vulcanologia (INGV), which enable detailed tracking of magma dynamics and early warning of eruptive activity. In Spain, the Cumbre Vieja volcano on La Palma (Canary Islands) demonstrated the effectiveness of such monitoring during its 2021 eruption, when thousands of seismic events were detected in the weeks preceding the crisis [36, 37].

At least seven stratovolcanoes in mainland Mexico have exhibited eruptive activity within the past five centuries. These include Sangangüey, Ceboruco, Colima, Paricutín–Tancítaro, Popocatépetl, Citlaltépetl, San Martín Tuxtla, Chichón, and Tacaná volcanoes. The most hazardous eruptions recorded during this period occurred at Ceboruco (1870–1875), Chichón (1982), and Colima (1913).

Colima Volcano was the first to be instrumented, with monitoring efforts initiated in 1989 [38]. By February 14, 1991, three seismic stations were operational, enabling the forecasting of the 1991 eruptive event [39] and subsequent eruptive processes [40]. Currently, all of the aforementioned volcanoes are instrumented, with the sole exception of Ceboruco [41].

These results confirm the need for a permanent seismic network on and around this active volcano to monitor the volcanic system in real time, enabling short-term forecasting. The seismic data provided by a permanent seismic network will enable the application of more advanced methods for identifying microearthquakes, tomographic methods to improve understanding of internal volcanic structures, and defining variations in seismic velocities. These studies complement the work presented here and clarify the hypotheses that emerge. These international examples highlight the global importance of maintaining robust seismic networks, which advance volcanological research and constitute a cornerstone of disaster risk reduction (DRR) strategies.

The findings presented herein underscore the critical imperative for the deployment of a permanent seismic monitoring network encompassing the Ceboruco volcanic complex. Such infrastructure is essential for continuous, real-time surveillance, which is a prerequisite for effective short-term eruption forecasting. The acquisition of a continuous, high-fidelity dataset would facilitate the implementation of sophisticated analytical techniques, including matched-filter detection for microseismicity and seismic tomographic inversion. These methods are necessary to resolve the internal subsurface architecture and to quantify spatiotemporal variations in seismic-velocity structures. Future investigations utilizing these advanced metrics would complement the baseline characterization established in this study and rigorously test the hypotheses proposed regarding magmatic reactivation. Aligning with international best practices, maintaining robust instrumental networks is not only fundamental to advancing volcanological knowledge but also a cornerstone of effective DRR and hazard mitigation strategies.

Given the historical record of high-magnitude explosive episodes at Ceboruco Volcano, juxtaposed with the region’s substantial demographic expansion and infrastructure development, it is evident that societal vulnerability and cumulative risk exposure have escalated. Consequently, establishing a robust geophysical baseline is imperative to ensure the accurate detection and interpretation of potential precursory anomalies.

Acknowledgments

Projects that funded this research: CONACyT 2007-01-0080034; Project CONACyT-Fondo Mixto Jalisco (FOMIXJal) 2008-96567 (2009); Project CONACyT–FOMIXJal 2008–96539 (2009); Project CONACyT-FOMIXJal 2010–149245 (2010); Project CONACyT [398-T9402] and UdeG internal projects; Centro Mexicano de Innovación en Energía – Geotérmica (CeMIE-Geo), P24, Project Secretaria de Energía – Consejo Nacional de Ciencia y Tecnología MEXICO (SENERCONACyT) 201301-207032; partially supported under the US Department of Energy Contract 89233218CNA000001.

Conflict of Interest

The authors declare no conflict of interest.