This research covers the study of the first month of data from a local seismological network deployed after the MW 6.4, January 18, 2021, San Juan earthquake, over the central Chilean-Pampean flat slab. Almost 1000 seismic events were detected in the first month after the MW 6.4 earthquake occurred, with ML magnitude ranging from −0.2 to 4.3. The majority of the seismic events were relocated at a depth of around 14 km and distributed in a northeast–southwest direction. The focal mechanism solutions obtained represent the rupture main process with one of their nodal planes according to the epicentral distribution, from which we can define the rupture dip to 61°. On the other hand, quickly after the MW 6.4 earthquake occurred, a nearby region at shallower depths became activated, from which Differential Synthetic Aperture Radar Interferometry results show vertical movements. In the epicentral area and concordance with the northeast–southwest rupture region, we found a low of the analytical signal which extends in the same strike by approximately 80 km to the most populated region in the San Juan Province. Considering the closeness of the greatest earthquakes that shocked the eastern sector of the San Juan Province to the subducted Juan Fernandez Ridge track, and the direction of the rupture process, we infer its influence as one of the significant factors that contributes to the high seismic activity in this region.

The western margin of the South American plate constitutes one of the most tectonically active regions in the world associated with destructive subduction-related seismicity and orogenic (Andean)-related seismicity (Figure 1). In addition, this seismicity extends further into the plate interior where the Nazca and Cocos plates subduct subhorizontally beneath the South American plate, such as in the case of the Chilean-Pampean flat slab between ~30° and 33°S (Figure 1). The absence of a quaternary volcanic arc in these shallow subduction settings is associated with cooler lithospheric conditions and progressive dehydration of the subducted slab that leads to higher seismic activity [1]. This catalog starts at the end of the 19th century, in 1894, with an earthquake with a maximum MM (Modified Mercalli) intensity of IX, representing one of the greatest earthquakes in Argentina and particularly in the San Juan Province (Instituto Nacional de PREvención Sísmica “INPRES” historic register). After that, with similar maximum intensity, on January 15, 1944, an earthquake of MW 7 [2] shook the region causing numerous fatalities and economic losses. After this, the greatest earthquake in Argentina and the San Juan Province was registered in 1977 with Ms 7.4 [3]. Recently, on January 18, 2021, at 23:46 local time, an earthquake with MW 6.4 (INPRES Catalog, NEIC-USGS Catalog) alarmed the population, causing socioeconomic damages mainly in the Pocito Department in the San Juan Precordillera and minor damages in the western Rivadavia and Sarmiento departments. Also, secondary processes were observed, such as rockfall and decrease or increase in water flows in downstream slopes, among others [4]. In Figure 2, we plot more than 20 years of crustal seismicity (INPRES Catalog) and the greatest earthquakes which occurred around the January 18, 2021, earthquake. In particular, the 1944 and 1952 hypocenters have high uncertainties in their locations considering the lack of seismological stations at that time (for details, see discrepancies in the earthquake locations for the 1944 and 1952 San Juan earthquakes from different sources in Table 2 of [2]).

The January 18, 2021, earthquake, its replicas, and triggered seismicity (with the first reported [from INPRES catalog] seismic event hours later the main shock) become an exceptional opportunity to study the active structure of the region, considering that from 2011 to 2020, a total of 115 crustal events were reported with magnitude values between 2.5 and 4.7 (Figure 3(a)), while after this event, a total of 252 seismic events were reported solely in the first week (Figure 3(b)) (both groups of events were taken from online INPRES catalog).

To register the aftershocks and the activated seismicity associated with this earthquake, the Volponi Institute (IGSV), from San Juan University, deployed a broadband seismological temporal network (Figure 3(c)) which operated with nine stations, 9 days after the main shock [5]. In addition, magnetic data from the global EMAG2v2 model complemented the seismological results and achieved a more robust interpretation of the structure at depth.

1.1. Data and Methodology

Immediately after the MW 6.4 January 18, 2021, earthquake, a seismological campaign was organized. The completed experiment of the broadband seismological network operated from January 27 up to July 4, 2021. The seismometers deployed are Trillium 120 PA, Trillium Compact, and Guralp 6T, and the seismographs are Centaur, Taurus, Q330, and Minimus. All the stations registered data continuously at 100 samples per second. Data were organized by an hour record length in the Seisan [6] database. In order to find the most possible quantity of seismic events, a visual inspection was carried out instead of using an automated detection software. P and S wave arrivals were identified using a multitrace view in Seisan. We located seismicity using Hypocenter [7] software, which located events individually, using a one-dimensional model from Villegas [8]. These procedures were done for all the recognized seismic events, although in this work, only crustal events (intermediate-depth earthquakes are not included) are presented. The local magnitude (ML) for those seismic events was calculated in the Seisan platform and by semiautomatic mode from the S wave maximum amplitude (then averaged) from each vertical component [9]. Then, in the second stage, those locations were improved using HypoDD [10], which is a double-difference (DD) earthquake location software. HypoDD uses the benefit that if the hypocentral distance between two seismic events is small, in contrast with the distance between the station and seismic events, and the scale length of the velocity structure heterogeneity, then for the zone around the source and the current station, the ray paths will be similar nearly the whole travel [11]. So, the travel times discrepancy for two seismic events found at one station could be the product of the spatial offset between the events, considering the common errors cancel (e.g. an unknown structure near the station) in the double difference procedure [11]. HypoDD uses a combination of catalog (P and S phases) and cross-correlation (waveforms) data. Before the relocation software is run, data preprocessing is made using phd2dt [11] which converts the phase data and/or waveform data to travel time differences for pairs of seismic events at common stations. This preprocessing is essential to enhance the connection between the seismic events and avoid repetition in the dataset. After that procedure, 967 seismic events were selected from the total database (979 seismic events) to be relocated. HypoDD in its procedure allows the use of different parameter configurations, like admitting only one type of data (waveform or catalog) or admitting both types of data (waveform and catalog), between others. One of the most crucial is the definition of dynamic weighting during the relocation process. In this work, for the weighting scheme, we use (for the first 30 iterations) low and high weights for the cross-correlation and catalog data, respectively, to let the catalog data generate a first global view. Then, in the next 30 iterations, we continued with low weight for cross-correlation data, but we re-weighted the catalog data to reject events with large misfits in seconds or great distance separation between events. Throughout, the subsequent 30 iterations we assign low and high weights for catalog and cross-correlation data, respectively, to enhance events whose waveforms correlate. Also, only event pairs with separations smaller than 2 km are allowed in the cross-correlation data. Finally, in the last 30 iterations, only events with separation lengths lower than 500 m, are used in the cross-correlation data, to conclude the definition of possible very local structure.

Using earthquake relocation obtained in the previous named stage, we select events with a magnitude greater than 3.0 (to ensure a high signal-to-noise ratio) to obtain focal mechanisms solution from the P first motion polarities. To obtain reliable focal mechanism solutions, we complemented our database with available online waveform data (FDSN network) and data from IGSV stations located in Mendoza and Neuquén Province. The polarities were identified with up or down for each station dividing into four quadrants, two compressional and two dilatational. For the calculation, we used the FOCMEC [12, 13] program which realized an effective search of the focal sphere and informed acceptable solutions. In this work, for the degree increment search, we use 5° to find solutions and a minimum of 15 first-motion polarity readings.

For the magnetic data, we used the global EMAG2v2 (www.ncei.noaa.gov) which is a global Earth Magnetic Anomaly Grid compiled from satellite, ship, and airborne magnetic measurements. The resolution is 2 arc-minutes (~3.7 km), and the altitude is 4 km above the geoid. We obtained an EMG2v2 data grid for the study area and we applied the analytical signal [14, 15,16] to analyze the geometry of the geological structures (medium to long wavelength).

The aftershocks have magnitudes from −0.2 to 4.3 Ml, with a median magnitude of 1.0 Ml; and depths from 6 to 30.4 km, with a median of 18.8 km (Figure 4).

In general, most seismic events (approximately 60%) are located between 15 and 20 km, in minor quantity (approximately 30%) between 20 and 30 km depth, and the rest (approximately 10%) at shallower depths, between 5 and 15 km depth. The location parameters present similar errors in their determination; however, the lowest error is in the longitude with a median error of 1.2 km compared with the 1.7 and 1.8 km in latitude error and depth error, respectively. The latter is due to a geometrically unbalanced network (better coverage is at longitude instead of latitude). The epicenter distribution shows a mainly northeast–southwest tendency (Figure 4), which follows one of the nodal planes obtained for the main shock [17, 18]. On the other hand, after the earthquake’s relocation, 935 seismic events were concentrated in a smaller epicentral region (Figure 5(a)) and events moved to shallower depths, with a median depth of 18.8to 14.8 km, after the relocation (Figure 5(b)).

In the plot of Figure 5(b), it is clear how the group of seismic events moves to a shallower depth and concentrates in a narrower region, after the second location stage. Then, even though seismic events in the first location stage using Hypocenter software [7] located entirely in the Zonda and Sarmiento Department (see gray circles in Figure 5), after the second stage using HypoDD relocation software [11], they concentrated on the Zonda Department (see blue circles in Figure 5(a)). Besides, the earthquake relocation using the double-difference algorithm, allowed us to divide the registered seismicity (for interpretation purposes) after the MW 6.4, 2021 earthquake, into two principal groups, both of which presented northeast to southwest orientation (Figure 6). The first and more numerous group, located principally between 14 and 20 km depth, shows the rupture process related to the main shock. The second shallower group, between about 9 and 12 km depth, was concentrated in a more restricted area, related to a region activated after the main shock.

The obtained focal mechanism solutions (using FOCMEC software) are plotted in Figure 6(a) and detailed in Table 1. The focal mechanisms 1 and 2 are similar to the main shock, denoting a strike-slip solution with approximately northeast to southwest and northwest to southeast nodal planes. In the case of focal mechanism solution 1, the nodal plane northwest to southeast is near the vertical, so a little variation in its inclination allows two possible solutions: a strike-slip solution with a thrust component (named 1-A) and a strike-slip solution with a normal component (named as 1-B). Those solutions are plotted in black in Figure 6 (more likely, with a thrust component similar to the main shock) and in gray color (less probable, with a normal component). On the other hand, focal mechanism 3, related to a shallower event (see focal depths in Table 1) presents two possible solutions, a normal solution with a strike-slip component (named 3-A) and a strike-slip solution with a normal component (named 3-B). The nodal planes (for the three seismic events), plotted in Figure 6 and presented in Table 1, are the media solutions of each group of solutions found for each event, for details see supplementary material which shows the complete results (the set of solutions for each seismic event: see online Supplementary Material Figure S1-S3).

To define the rupture process related to the main shock of the MW 6.4, January 18, 2021, earthquake, we plotted in Figure 7(a) more restrictive (as regards uncertainties) dataset in which only seismic events with an azimuthal gap less than 130° are indicated (a smaller azimutal gap means smaller hypocenter errors). Also, the focal mechanism 1 (solution A) and the focal mechanism 2, together with focal mechanism solutions for the main shock [17, 18], are plotted, to visualize whether a nodal plane coincides with the seismic epicentral distribution. Then the nodal planes with the northeast–southwest direction can be associated with the main fault plane of the earthquake of MW 6.4 on January 18, 2021. The last is in accord with the results from [17], in which the authors interpreted the nodal planes with an azimuth, dip, and rake of 219°, 78° (to the northwest), and 172°, respectively.

Additionally, we present a group of recorded events located at shallower depths mostly between 9 and 12 km (which are plotted in dark red color in Figure 6(a) and plotted alone in Figure 8). We interpret that this group of events is not directly associated with the main rupture zone of the MW 6.4, January 18, 2021, earthquake, but more likely triggered by this. Particularly, in this region, from Differential Synthetic Aperture Radar Interferometry, surface displacements up to ~4.5 cm in the LOS direction (descending orbit) using time series data from January 2021 were determined [19]. Unfortunately, the group of events reported in this work does not have a reliable focal mechanism solution, since the unique focal mechanism solution (event 3) at the same epicentral area occurred approximately 5 months later (June 24), with a focal depth of 4.8 km. Therefore, it is not possible to reliably determine if the focal mechanism of the June 24, 2021, event is representative of this group of late events (9 and 12 km in depth, distributed by 2 km in the northeast to southwest direction).

A direct and preliminary interpretation would be that this group of shallower events (9 to 12 km in depth) could have been related to the Maradona fault, a well-known, considered active thrust fault associated with neotectonic indicators (Figure 8) [20]. However, again, we prefer to be cautious about this potential relationship since this region, which started to activate just some hours later than the main shock of MW 6.4, with magnitude ML 3.8 (INPRES Catalog), continued its activity up to June and with a shallower event of ML 3.3 that had extensional movements (see Table 1 and Figure 6).

We can speculate then that this measured normal movement was in response to the “re-accommodation” of the vertical displacement found (after the 6.4 earthquake) by [19] which would have begun with the ML 3.8 earthquake reported by INPRES.

What is clear is that the main group of events oriented in a northeast to southwest direction and associated with the main shock (MW 6.4 on January 18, 2021) is related to a strike-slip fault, solved from the focal mechanism (events 1 and 2 and the main shock, Figure 7). The epicenters are distributed along 19 km (Figure 9) (seismic events as white circles with an azimuthal gap of less than 130°) in accord with the empirical relation of Well and Coppersmith [21] for an earthquake of magnitude MW 6.4.

Therefore, from the evidence provided in this work, we can define the fault which caused the earthquake of MW 6.4 on January 18, 2021, as a right-lateral strike-slip fault. Most of the events are located around 14 km depth with a decreasing number of events around 20 km, and a median depth of 15 km which is deeper and shallower than the depths obtained for the main shock by Sanchez [17] and Amirati [18], respectively. From this, we remark on the importance of a quick seismological local network deployment (with good coverage around the epicentral area and at least one station near the epicenter) after a moderate magnitude earthquake occurs, at least during the first month of aftershocks. The lack of surface expression of the MW 6.4 earthquake is in accord with the depth of the rupture zone. However, the CORD station after the January 18, 2021, earthquake provides evidence of surficial changes.

Additionally, we used the analytical signal from EMAG2 to map the basement structure associated with the seismic series of January–February 2021. From this dataset, a series of magnetic anomalies with northeast-southwest regional structure is observed (without significant surficial expression). This mapped 80 km long basement structure projects from the Central Precordillera into the more populated area and with higher economic significance in the Pocito and Rawson departments in the eastern San Juan Province (Figure 10). If this interpretation is correct, the seismic series registered from January to February 2021 would have ruptured only the southwesternmost section of a much longer basement structure crossing obliquely the Precordillera fold and thrust belt at depth (Figure 10(b)). Since historic earthquakes have high uncertainties in their location parameters, their re-ubication could be related to deep basement structures decoupled from the thin-skinned shallower structure, such as could be the case of the Ms 6.8 June 11, 1952 [2], San Juan earthquake, preliminarily related to the reverse north-south La Rinconada fault [22, 23] but with no clear origin, that could be related to this region of low analytical signal (10 km to the south approximately).

Finally, in a more regional analysis, the orientation of the analyzed seismic series and the determined basement structure is similar and lies near the track of the Juan Fernández ridge which is subducted at these latitudes in association with the southern Chilean-Pampean flat slab and is associated with anomalous seismic energy release (Figure 10). The most important historical events such as 1944, 1952, and 1977 events occurred directly above or close to the edges of the Juan Fernández ridge path beneath this section of the Chilean-Pampean flat subduction zone. A direct relationship is visualized between these occurrences and higher seismic energy releases associated with the subduction of this volcanic relief. Unlike the interpretation of Sanchez et al. [17], in which they interpret the January 18, 2021, MW 6.4 earthquake and the aftershocks, as due to a basement structure reactivation associated with a weakness zone located in the basement of the Precordillera sedimentary series, we do not discard that these events could have been linked to primary structures influenced by the passage of the Juan Fernández ridge at depth.

This work allowed us to characterize the rupture process of the MW 6.4 January 18, 2021, earthquake, with a new local network set deployed after the main shock, limiting its subsurface activation region to a depth of around 14 km, different than previous estimates. Also, the highly accurate relocation process allowed us to highlight a shallower independent seismic cluster probably triggered by the main shock and coincident with the region where vertical displacement was observed using satellite interferometry [19]. Evidence of a deep basement structure is determined using the EMAG2 dataset, as the source of these events, obliquely crossing the Precordillera fold and thrust belt and decoupled from the north-south set of thin-skinned Neogene to neotectonic structures. We finally remark on the link between the most active region of San Juan Province located on the Chilean-Pampean flat subduction zone with the track of the subducted Juan Fernandez Ridge at depth and the activity of the basement structure responsible for the studied seismicity.

No additional data are available.

The authors declare that they have no conflicts of interest.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

The authors acknowledge the two anonymous Reviewers and Editor who allowed improving this work. They are grateful to the National San Juan University and to CONICET. We would like to extend our gratitude to the facilities of IRIS Data Services, and specifically the IRIS Data Management Center, which were used for access to part of the waveforms used in this study. IRIS Data Services are funded through the Seismological Facilities for the Advancement of Geoscience (SAGE) Award of the National Science Foundation under Cooperative Support Agreement EAR-1851048. The authors are grateful to the Instituto Nacional de Prevención Sísmica (INPRES) from the online Catalog which allows the authors to show a long period of crustal seismicity. Authors acknowledge the use of the GMT-mapping software of Wessel et al. (2019) https://www.generic-mapping-tools.org/ and the QGIS software https://qgis.org/es/site/.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Set of focal mechanism solutions obtain for the seismic events of: January 29, 2021 Ml 3.3, January 31, 2021 Ml 4.3 and June 24, 2021 Ml 3.3. In each case, the waveform of the first P wave arrivals at each station is picked and classified according to whether it is distensive or compressive.

Supplementary data

Supplementary Material 1
- docx file