Introduction and Tectonic Setting

From December 9^th, 2011 to February 28^th, 2012, an earthquake swarm took place in Central Costa Rica near the town of Tobosi, ~14 km southeast from San Jose, the capital city of the country. On a regional scale, this earthquake swarm is located on a portion of the Central Costa Rican Deformed Belt (CCRDB), an ~100-km broad zone of deformation that represents the western boundary between the Caribbean Plate and the Panama Microplate (i.e., Marshall et al., 2000; Montero,2001, Fig 1A). Regionally, the Cocos Plate subducts underneath both the Caribbean Plate and the Panama Microplate. The CCRDB connects the Middle American Trench on the Pacific side to the North Panama Deformed Belt on the Caribbean (Fig. 1A). The CCRDB includes active faults and folds of diverse geometry, but in general three different domains can be recognized (i.e., Marshall et al., 2000; Montero, 2001). In the western domain, located in the Pacific forearc, faults are mainly of transtensive type striking northeast-southwest. In the central domain, located approximately in the inner volcanic arc, faults are mainly of strike-slip type with right lateral displacement along faults striking northwest-southeast and left lateral displacement along faults striking northeast-southwest. The eastern domain, in the Caribbean back arc, is dominated by transpression faults. The origin of the CCRDB could be related to the collision of the buoyant over-thickened Cocos Ridge, which drives both shortening and lateral scape of crustal blocks over a broad zone (i.e., Lewis et al., 2008; Montero et al., 2013).

In a more local scale, the Tobosi earthquake swarm occurs between two major well-known left-lateral strike-slip fault systems: Aguacaliente and Navarro (i.e., Montero, 2001; Fernández & Montero, 2002, Fig. 1B). Several historical damaging earthquakes have impacted Central Costa Rica, including the magnitude (Ms) 6.1 Cartago (Sta. Monica) earthquake on May 4th, 1910, which has been the most destructive earthquake in the history of Costa Rica (i.e., Montero & Miyamura, 1981, Fig. 1B).

The National Seismological Network of Costa Rica (RSN) is one of the entities that monitors the seismicity of the country. The RSN derives from an agreement between the Seismology and Volcanology Division of the School of Geology at the University of Costa Rica (UCR) and the Seismology Unit of the National Electricity Institute (Instituto Costarricense de Electicidad -ICE). The Seismology Laboratory at the UCR was founded in 1973 with a data catalogue that goes from 1976 to the present. During 2011 and 2012, the RSN ran a network of 62 permanent stations.

This paper describes the Tobosi earthquake swarm in terms of its time and space distribution and determines its tectonic source. These goals are accomplished by relocating earthquakes and calculating focal mechanisms using data recorded by the RSN and by analyzing the tectonic geo-morphology of the region. Given the history of destructive earthquakes in central Costa Rica, studies of active faults in this area are relevant for determining seismic hazards for this region, where most of the population lives and most of the economic and industrial activities are concentrated.

Methodology

Earthquakes were selected from the RSN data catalogue. The earthquake search was performed using the following criteria: earthquakes occurring from December 9th, 2011 to February 28th, 2012 located in the Tobosi region, with at least ten P-wave phases and one S-wave observation, with good seismic station coverage (i.e., GAP < 180), and with a root mean square (RMS) of less than 0.5 s in the initial location. These earthquakes were recorded by 37 RSN stations, including sites from the permanent RSN network as well as temporal stations from ICE (figures 1A and 2A). These stations include both short period (Lennartz 3D Lite) and broadband sensors (Guralp). "La Lucha" (LCR2, Fig. 2A) was the nearest station to the earthquake swarm, located at ~8 km. On the other hand, the farthest station was located at the Rincon de la Vieja volcano (station GPS3), 189 km away from the epicenters. Earthquakes were clearly recorded at stations located less than 70 km from the source (Fig. 2A).

The software package used for earthquake processing and analysis was SeisAn (Ottemoller et al., 2011), which is an open-source software commonly used in Central American Seismology institutions. The earthquake database managing and picking was performed by using the programs "eev" and "mulplt", respectively, both from the SeisAn package. The time window used for P-wave and S-wave picking was 5 and 10 s, respectively. P-wave arrival was usually very clear for the nearest stations to the source (Fig. 2B). A five-level weighting scheme was used for categorizing the uncertainty of the picks. Level zero corresponds to 100% pick confidence while level four corresponds to 100% pick uncertainty.

Earthquake locations were found using a seven-layer linear 1D model, used for routine locations at RSN, also based on the model proposed by Matumoto et al. (1977). The response files for each of the stations were built with an open-source software called PDCC (Portable Data Control Center) from Casey (2012). Additionally, first motion focal mechanisms were calculated using the FOCMEC (Focal Mechanisms Determinations; Snoke, 1984) program. Finally, field trips were conducted in the region to characterize the local Tectonic Geomorphology and to complement the earthquake analysis.

The Tobosi Swarm and Historical Earthquakes

Examining earthquake spatial and temporal distribution is fundamental to understand the physics of the earthquake source and cycle (i.e. Eneva & Hamburger, 1989). An earthquake swarm is defined as an increment of earthquakes in a limited area within a limited period of time, from several days, weeks or even months, and without connection to a main shock (i.e., Hill, 1977; USGS, 2004).

In the vicinity of the town of Tobosi, 22 earthquakes occurred from December 9th, 2011 to February 28^th, 2012. Twenty of these earthquakes occurred within a time frame of 13 days.

The swarm initiated in December 9^th, 2011 with a 3.3 Mw earthquake and a depth of less than 1 km. Most of the events occurred between December 22^nd and 27^th, 2011. The dates with most events were December 23^rd and 26^th, with five and six of them, respectively. The last event occurred on February 28^th, 2012 with a magnitude Mw of 3.3 and depth of 5.8 km. This event is not included in the date histogram (Fig. 3A) in order to allow a better visualization of the scale of this histogram.

Earthquake moment magnitudes (Mw) for the Tobosi swarm ranged between 2.4 and 3.9, with the strongest earthquake occurring on December 26th, 2011. Most of the earthquakes (20) were between 2.6 and 3.5 Mw (Fig. 3B). Hypocenters ranged between 0.6 and 8.2 km in depth, with an average of 2.4 km (Fig. 3C, Table 1). Most of these earthquakes were felt at Tobosi and the largest ones were felt as far as in the cities of Cartago and San Jose.

There have been several historical earthquakes in the vicinity of the Tobosi swarm (Montero & Miyamura, 1981; Ambraseys & Adams, 2001; Fig. 1B). The closest significant earthquakes have been associated with the Aguacaliente Fault. These took place on April 13th, 1910 at 7:05 UTC (Coordinated Universal Time), with a magnitude of 5.6 (i.e., Peraldo & Montero, 2010), and on May 4^th, 1910 at 18:47 UTC south of the city of Cartago (i.e., Montero & Miyamura, 1981) with a magnitude Ms of 6.1. These earthquakes caused significant damage in the central part of Costa Rica, especially the Cartago earthquake, considered the deadliest natural disaster in the history of Costa Rica (Montero & Miyamura, 1981; Montero, 2010). Two other significant earthquakes struck the area; the first one on February 21st, 1912 near the town of Tres Rios, with a magnitude calculated based on intensity data (Mi) of 5.0 (Morales, 1985) and the second one, on August 22nd, 1951, near the city of Paraiso, with a magnitude Ms 5.4 (Ambraseys & Adams, 2001). The latter event might be associated with the Navarro Fault (Montero, 2001).

The Tobosi Transtension

Based on tectonic geomorphology, a transten-sional region was found near the town of Tobosi, which is referred to as the Tobosi Transtension (Fig. 4) in this paper. This structure is located between two major left-lateral fault systems mentioned above: the Aguacaliente Fault to the north (Woodward & Clyde, 1993; Montero et al., 2005; Montero & Kruse, 2006; Montero et al., 2013; Alonso-Henar et al., 2013) and the Navarro Fault to the south (i.e. Berrangé & Whitaker, 1977; Aguilar, 1984; Geomatrix, 1994; Montero et al., 1998; Linkimer, 2003; figures 4 and 1B). The Aguacaliente Fault is located about 3 km north of the Tobosi earthquake swarm. This active fault is 15-18 km long and is a seismogenic source capable of producing large earthquakes (Mw 6.6-6.9) with an estimated recurrence interval of about 500 years (Alonso-Henar et al., 2013). The Navarro Fault is located 7 km south of the Tobosi earthquake swarm. The accumulative longitude of different segments of the Navarro fault system is about 54 km and it might be capable of generating earthquakes of Mw 6.7 (Linkimer, 2003).

The Tobosi Transtension structure includes three fault traces: Tobosi, Tablon, and Alumbre, each of them displaying a left-lateral strike-slip displacement with a normal component. The Tobosi Fault bounds the Tobosi Transtension structure to the northwest. It strikes ENE-WSW and is about 10 km long. Initially, this fault was named Tablon Fault by Woodward & Clyde (1993) and interpreted by these authors as a potentially active fault. Considering its proximity to the town of Tobosi, it will be called the Tobosi Fault in this study. The fault trace proposed by Woodward & Clyde (1993) was also extended to the southwest, along the linear river valley of the Molina Creek (Fig. 4). To the southwest of Tobosi, this fault displays an east-faceted scarp and the Purires River flows in a parallel direction to it. A left-lateral displacement with a normal component was interpreted for this fault.

The Tablon Fault is the central trace of the Tobosi Transtension structure. This fault is about 8 km long and displays a left-lateral strike-slip displacement along the segment where the fault strikes east-west; and oblique (normal-left lateral) where the fault strikes ENE-WSW. The main geomorphological features of this fault are faceted scarps, linear riverbed valleys and deflected river drains. At the intersection between the Tablon fault and the Purires River (site 1 in Fig. 4), a fault scarp deflects to the left of the river.

The Alumbre Fault (Fig. 4) bounds the Tobosi Transtension structure to the south. It strikes east-west, is about 8 km long and displays a left-lateral strike-slip movement. The geomorphological evidences for this fault include linear riverbed valleys, mountain saddles, faceted scarps on both sides of the fault and deflected hills and mountain divides.

Focal Mechanisms Solutions and Origin

First-motion focal mechanisms were calculated for the events with more than seven polarities (Fig. 5). The software used to compute the focal mechanisms is called FOCMEC (Focal Mechanisms Determinations; Snoke, 1984). The search for nodal planes was performed every five degrees in the focal sphere. The quality of resulting focal mechanisms was categorized in two groups: A and B (Fig. 6). Quality A (QA) corresponds to the best quality and has at least 13 polarities mapped with no polarity errors in the focal sphere and sometimes with only one possible solution (Fig. 5). Quality B (QB) corresponds to focal mechanisms with at least seven polarity observations mapped with no polarity errors in the focal sphere. Fourteen focal mechanisms were computed: seven of QA and seven of QB (Fig. 6, Table 2).

A strong trend on the polarity observations was found for most of the focal mechanisms for both QA and QB. In QB focal mechanisms, the number of polarity observations allowed many possible solutions, so that even when the map of polarity observations in the focal sphere was very similar to those of QA, they could not be considered as high-quality solutions and only QA focal mechanisms were considered for the tectonic interpretation (Fig. 7). The trend of polarity observations noted for QB focal mechanisms suggests that most of them might have similar solutions for most QA. A possible solution for each QB was found in Fig. 6.

QA focal mechanism solutions are mostly strike-slip faults. The polarity observations and nodal plane solutions of six events are nearly identical (events number 1, 2, 5, 7, 12 and 14, Table 2, Fig. 7). These solutions correspond to a left-lateral strike-slip solution (rake angle varies from 3 to 12°) for a nodal plane striking from 254 to 264 and dipping from 71° to 80°, and to a right-lateral strike-slip solution (rake angle varies from 161° to 170°) for a nodal plane striking from 162 to 172 and dipping from 78° to 87°. Pressure and Tension axis for this set of QA focal mechanisms have a strike of 28°-38° and 118°-128°, respectively, and a dip angle of 1°-8° and 15°-19°, respectively (Table 2).

The main cluster of epicenters for the Tobosi swarm is aligned ENE-WSW and covers a region of approximately 3 km long and 1 km wide to the west of the town of Tobosi (Fig. 1B). The epicenters are also located at the eastern portion of the Tobosi Fault, where it has a strike of ENE-WSW. The left-lateral strike-slip solution for nodal planes striking ENE-WSW from the QA focal mechanisms seems to agree very well with the Tectonic Geomorphology observations for the region (Fig. 7). However, a discrepancy was noticed in the dip angle. Based on the Geomorphology, the Tobosi Fault may be interpreted as a nearly vertical fault inclined towards the SSE. On the other hand, the nodal plane solutions for focal mechanisms are nearly vertical (average of 77°), but dipping towards the NNW. Uncertainties in the search for nodal planes in the focal sphere for focal mechanisms may explain this disagreement, especially because the dip angle is almost vertical. An outcrop exposing the Tobosi fault plane is needed to more accurately determine the dip direction of this fault.

Based on the agreement between the fault trace location and the earthquake swarm locations, and on the left-lateral strike-slip solution for nodal planes striking ENE-WSW from the QA focal mechanisms, it is suggested that the Tobosi Fault was the source of the Tobosi swarm (Fig. 7).

Only one QA focal mechanism displays a different solution compared to the main trend (event number 6 in Table 2). This was the solution for an earthquake that happened in December 23rd, 2011, located to the south of the main cluster of earthquakes (Fig. 7). This event shows a right-lateral strike-slip solution (rake = -170°) on a plane striking 306° and dipping 56° and an oblique (normal-left lateral) solution (rake = -33°) for a nodal plane striking 210° and dipping 81°. This earthquake might have been produced by a different fault. Based on the Geomorphology where this epicenter is located, no clear indication of faulting was found, so it is not possible to determine which nodal plane may correspond to the active fault that originated this particular event. Interestingly, another fault showed seismic activity during the swarm occurring mainly at the north-easternmost portion of the Tobosi Fault.

Conclusions

In the vicinity of the town of Tobosi, an earthquake swarm of 22 events occurred from December 9th, 2011 to February 28th, 2012. Twenty of these earthquakes occurred on a 13-day time frame. Earthquake magnitudes (Mw) varied between 2.4 and 3.9 and focal depths between 0.6 and 8.2 km. This earthquake swarm occurred on a portion of the CCRDB where faults are mainly of strike-slip type with right-lateral displacement along faults striking northwest-southeast and left-lateral displacement along faults striking northeast-southwest.

Based on Tectonic Geomorphology, a trans-tensional structure near the town of Tobosi was determined, referred in this study as the Tobosi Transtension. This structure is located between two major left-lateral fault systems: Aguacaliente and Navarro. The Tobosi Transtension structure is composed of three faults, named Tobosi, Tablon, and Alumbre. Each of these faults displays geo-morphological features that suggest left-lateral strike-slip displacements with a normal component. Some geomorphological evidences for these faults include linear riverbeds valleys, mountain saddles, and faceted scarps.

A strong trend on the polarity observations for focal mechanisms was found for most earthquakes analyzed. High-quality focal mechanism solutions display mostly strike-slip faults. These solutions correspond to a left-lateral strike-slip displacement on a nodal plane striking 254 to 264 and dipping from 71° to 80°, and to a right-lateral strike-slip displacement on a nodal plane striking 162 to 172 and dipping from 78° to 87°. The main cluster of epicenters occurs to the west of the town of Tobosi, is aligned ENE-WSW, and covers a region of approximately 3 km long and 1 km wide. The epicenters are also located at the northeastern portion of the Tobosi Fault where it strikes ENE-WSW. The left-lateral strike-slip solutions for the high-quality focal mechanisms seem to agree very well with the Tectonic Geomorphology observations for the Tobosi Fault in this region. Based on these observations, it is proposed that the Tobosi Fault was the source of the 2011-2012 earthquake swarm.

There have been at least five historical earthquakes in the vicinity of the Tobosi swarm. The most significant earthquakes have been associated with the Aguacaliente Fault (i.e., Cartago, 1910, Ms 6.1) and the Navarro Fault (Paraiso, 1951, Ms 5.4). Given this history of destructive earthquakes, studies of active faults are important for determining seismic hazards for the densely populated area of Central Costa Rica.

Acknowledgements

The collaboration between the co-authors was made possible through the research projects: 113-B3-129 "Estudio de la tectónica activa de las fallas del Cinturón Deformado del Centro de Costa Rica", 830-B1-504 "Estudio de la tectónica activa del sistema de fallas de Aguacaliente y Navarro del Valle Central mediante Técnicas Geodésicas y Paleosísmicas: contribución para la mejora de la evaluación de la amenaza sísmica", funded by the Science and Technology Ministry of Costa Rica and the University of Costa Rica and 11-B3-263 "Propiedades físicas de la listós-fera y sismogénesis en Costa Rica" of the Central American School of Geology at the University of Costa Rica. The waveforms used were acquired by the RSN agreement between the University of Costa Rica (UCR) and the Costa Rican Institute of Electricity.

References

AGUILAR, A., 1984: Sismicidad del 3 al 9 de setiembre de 1980 y su relación con la geología en la cabecera del río Navarro, Cartago.- 104 pp. Univ. de Costa Rica, San José [Thesis Lic.].

ALONSO-HENAR, J., MONTERO, W., MARTÍNEZ-DÍAS, J. J., ÁLVAREZ-GÓMEZ, J. A., INSUA-ARÉVALO, J.M., & ROJAS, W. 2013: The Aguacaliente Fault, Source of the Cartago 1910 destructive earthquake (Costa Rica).- Terra Nova., 2S: 368-373.

AMBRASEYS, N. N., & ADAMS, R. D., 2001, The Seismicity of Central America. A Descriptive Catalogue 1898-1993.- 309 pp. Imperial College Press, London.

BERRANGÉ, J. P. & WHITTAKER, J. E., 1977: Reconnaissance Geology of the Tapantí Quadrangle, Talamaca Cordillera, Costa Rica.- 72 pp. Institute of Geological Sciences Overseas Division, Report 37, Londres.

CASEY, R., 2012: Portable Data Control Center (pdcc) v3.8 User Manual.- - http://www.iris.edu/pub/programs/pdcc/PDCC_3.8_User_Manual.pdf [Consulted: February 20, 2014.- IRIS Data Management Center].

ENEVA, M., & HAMBUGER, M., 1989: Spatial and Temporal Patterns of Earthquake Distribution in Soviet Central Asia: Application of Pair Analysis Statistics.- Bull Seism. Soc. Am. 79: 147S-1476.

FERNÁNDEZ, M. & MONTERO, W., 2002: Fallamiento y sismicidad del área entre Cartago y San José, valle Central de Costa Rica.- Rev. Geol. Amér. Central, 26: 2S-37.

GEOMATRIX CONSULTANTS, 1994: Informe final: acueducto de Orosi, sub-estudio de la vulnerabilidad sísmica de la conducción: El Llano a Tres Ríos, provincia de Cartago, Costa Rica.- 139 p. Geomatrix Consultants Report, San Francisco, CA [Int. report].

HILL, D. P., 1977: Model of earthquake swarms.- J. Geophys. Res. 82: 1347-13S2.

LEWIS, J. C., BOOZER, A. C., LÓPEZ, A. & MONTERO, W., 2008; Collision Versus Sliver Transport in the Hanging Wall at the Middle America Subduction Zone: Constraints from Background Seismicity in Central Costa Rica.

LINKIMER, L., 2003: Neotectónica del extremo oriental del Cinturón Deformado del Centro de Costa Rica.- 103 p. Univ. de Costa Rica, San José [Thesis Lic.].

MARSHALL, J. S., FISHER, D. M., & GARDNER, T. W., 2000: Central Costa Rica deformed belt: Kinematics of diffuse faulting across the western Panama block.- Tectonics, 19: 468-492.

MATUMOTO, T., LATHAM, G., OHTAKE, M., & UMAÑA, J., 1977: Crustal structure in Southern Central America.-, Bull Seism. Soc. Am. 67: 121-134

MONTERO, W., ROJAS, W. & LINKIMER, L., 2013: Neotectónica de las fallas Ochomogo y Capellades y su relación con el sistema de falla Aguacaliente, falda sur macizo Irazú-Turrialba, Costa Rica.-, Rev. Geol. Amér. Central 48: 119-139.

MONTERO, W. & KRUSE, S., 2006: Neotectónica y geofísica de la falla Aguacaliente en los valles de Coris y del Guarco.-, Rev. Geol. Amér. Central 34-35: 43-58.

MONTERO, W. & MIYAMURA, S., 1981: Distribución de intensidades y estimación de los parámetros focales de los terremotos de Cartago de 1910, Costa Rica, América Central.- Rev. Inst. Geogr. Nacional, Julio-Diciembre: 9-34.

MONTERO, W., 2001: Neotectónica de la región Central de Costa Rica: Frontera W de la microplaca de Panamá.-, Rev. Geol. Amér. Central 24: 29-56.

MONTERO, W., 2010: El terremoto de Cartago del 4 de mayo de 1910: Aspectos sismológicos y neotectónicos.- En: PERALDO, G. & ACEVEDO, B. (eds): Efemérides de la destrucción de Cartago cien años después (1910-2010).- Ediciones Perro Azul, San José: 37-47.

MONTERO, W., DENYER, P., BARQUERO, R., ALVARADO, G. E., COWAN, H., MACHETTE, M. N., HALLER, K. M. & DART, R. L., 1998: Map and Database of Quaternary Faults and Folds in Costa Rica and its Offshore Regions. 63 pp. U.S. Geological Survey [Open-File Report 98-481].

MONTERO, W., BARAHONA, M., ROJAS, W. & TAYLOR, M., 2005: Los sistemas de falla Aguacaliente y Río Azul y relevos compresivos asociados, valle Central de Costa Rica.-, Rev. Geol. Amér. Central 33: 7-27.

MORALES, L. D., 1985. Las zonas sísmicas de Costa Rica y alrededores.- Rev Geol. Amér. Central, 3: 69-101.

OTTEMÖLLER, L., VOSS, P. & HAVSKOV, J., 2011: SEISAN: The Earthquake Analysis Software for Windows, Solaris, LINUX, and MACOSX, version 9.0.1.- 361 pp. Univ. of Bergen, Bergen.

PERALDO, G. & MONTERO , W., 2010: Sismicidad anterior y posterior a los terremotos del 13 de abril y el 5 de mayo.- En PERALDO, G. & ACEBEDO, B. (eds): Efemérides de la destrucción de la ciudad de Cartago. Cien años después (19102010).- Ed. Perro Azul, San José: 23-35.

SNOKE, J. A., MUNSEY, J. W., TEAGUE, A. G. & BOLLINGER, G. A., 1984. A Program for Focal Mechanism Determination by Combined use of Polarity and Sv-P Amplitude Ratio Data. En: Earthquake Notes, 55: 15-20.

U. S. Geological Survey (USGS), 2004: Yellowstone Earthquake Swarms". United States Geological Survey.- - http://volca-noes.usgs.gov/volcanoes/Yellowstone/ye-lowstone_monitoring_50.html [Consulted: February 20, 2014].

WOODWARD & CLYDE, 1993: A Preliminary Evaluation of Earthquake and Volcanic Hazards Significant to the Major Population Centers of the Valle Central, Costa Rica.- 89 pp. Ret Corporation, San Francisco [Int. report].

IR