CRUSTAL THICKNESS ESTIMATION BENEATH THE NORTHERN ANDES (COLOMBIA) FROM TELESEISMIC RECEIVER FUNCTIONS Esteban Poveda (Servicio Geológico Colombiano and Universidad Nacional de Colombia at Bogotá) Gaspar Monsalve (Universidad Nacional de Colombia at Medellín) Carlos A. Vargas-Jiménez (Universidad Nacional de Colombia at Bogotá) Introduction Receiver functions analysis is a simple method of extracting constraints on crust and upper- mantle structure and discontinuities, beneath permanent and mobile seismic networks. The recent presence of Broadband stations in Colombian Territory, Operated by the National Seismological Network of Colombia (NSNC), has provided new data for calculation of receiver functions for crustal thickness estimation beneath 18 stations in the region. We show some of the results of the receiver function analysis for Moho depth, and crustal discontinuities and anisotropy beneath the eastern Cordillera of Colombia. . Broadband stations Since 2007, the National Seismological Network of Colombia has installed broadband stations in Colombian territory, to monitor the seismic activity of the country, to study the seismic hazards and for a better understanding of the seismotectonics of the region. Most of the information recorded at stations is at a sampling rate of 100 counts/seg and the data are transmitted via satellite. Figure 1 shows the broad band stations used for the receiver function analysis. The National Seismological Network of Colombia (NSNC) Data For this study we used teleseismic records from January 2008 through March 2012 for NSNC stations. For BOCO station, records were used between 1994 and 1996, period during which it was operated by IRIS. More than 400 events have been analyzed. Receiver functions were selected so that the adjustment percentage was greater than 80% Some receiver function examples H-k Stacking Receiver Functions beneath Bogota Figure 1. Map of broadband stations of the National Seismological Network of Colombia. In red, stations we show receiver functions for. Figure 2. a) Selected events for calculating receiver functions. Blue star represents the location of Bogota; b) histograms of back-azimuthal distribution of earthquakes used in this study Figure 3. Individual Receiver Functions for stations MON (a) and BRR (b), ordered by back-azimuth and calculated in the time domain. We show 79 traces for MON and 51 for BRR. Ps phase at about 3 seconds from direct P-arrival at MON and at about 5.8 seconds at BRR are interpreted as Moho conversions. This is consistent with a greater crustal thickness beneath the Middle Magdalena Valley (between Central and Eastern Cordillera) than beneath the Caribbean Coast. Figure 5. The stacking technique of Zhu and Kanamori (2000) is based on the conversion of the Receiver Functions amplitude from the time-domain into a system based on crustal thickness (H) and Vp/Vs ratio (k), by using the predicted times of 3 principal crustal reveberations: Ps, PpPs and PsPs/PpSs. Examples for stations MON (a) and BRR (b) are shown. Figure 6. a) Radial receiver functions ordered by back-azimuth at station BOCO. b) Sinthetic radial receiver functions at station BOCO. We ran a synthetic test with a crustal thickness of 58 km and an anisotropic layer 7 km thick at 25 km depth (from 25 to 32 km) with horizontal east oriented 10% anisotropy, using Frederiksen and Bostock code (2000). Polarity inversions are observed at approximately 1-4 sec in the radial receiver function. More work needs to be done in order to find the model that best fits the experimental receiver functions. Conclusions Ps-P times indicate that crustal thickness around 25 km beneath the Caribbean coast of Colombia. At station CAP2 (near the junction of Panama Block, North Andean Block and the caribbean Plate) crustal thickness appears to be around 36 km. For the Upper- Middle Magdalena Valley, between the Central and Eastern Cordilleras, the estimated crustal thickness is between 40 and 45 km. At some locations beneath the high elevations, including the high peaks of the Central Cordillera and the Eastern Cordillera Plateau, crustal thickness may exceed 50 km. Receiver functions at station FLO2, at the eastern foothills of the Colombian Andes, suggest a Moho depth of about 40 km. Receiver functions at stations on the Eastern Cordillera Plateau (including station BOCO, Figure 6) seem to show some signals of anisotropy in the mid-crust, probably due to the presence of a fold and thrust belt in the eastern Cordillera, resulting in preferential orientation of micas and amphiboles in the upper-mid crust, due to shear stresses. Receiver Functions at stations in the Southernmost areas of the Colombian Andes do not show a consistent Moho conversion and require more work on interpretation. Methodology EARTHQUAKE SEARCH • NEIC earthquake search • Epicentral distances between 29° to 90° • Magnitude of moment, Mw > 5.5 • Data extraction in format SEISAN of NSNC • Data extraction in format SAC of IRIS-WILBER network (station BOCO) SIGNAL PROCESSING • Conversion SEISAN to SAC. • Organization of data with respect to back- azimtuh • Correction for baseline, trend, 20 counts / sec • Picking P-arrival, cutting of the waveform: 20 seconds before P-arrival and 120 seconds after • Evaluation of signal quality RECEIVER FUNCTION CALCULATION AND INTERPRETATION • Deconvolution with Ligorria and Ammon algorithm (1999) • Obtaining Receiver Functions with Gaussian filters 0.5 s, 0.8 s, 1.0 s and 1.5 s wide. • Move -out correction (equalization for an epicentral distance of 67°) • Evaluation of Receiver Function quality • Preliminary Interpretation • Calculation of the mean Vp/Vs for each station using a modification of Wadati diagrams • H-k Stacking of Receiver Functions with Zhu and Kanamori (2000) algorithm. • Ray-theoretical modelling of teleseismic waves in dipping, anisotropic structures, using Frederiksen and Bostock (2000) code. a) b) 0 2 4 6 8 10 12 14 16 18 Time (s) a) b) 0 2 4 6 8 10 12 14 16 18 Time (s) Constraints on Vp/Vs a) b) Figure 4. Wadati diagrams were made for many earthquakes recorded at all stations, with epicentral distances of less than 160 Km and depths within 30 km so we obtain Pg phases in the crust. a) Diagram for station MON; b) diagram for station BRR. a) b) a) b) Crustal thickness beneath Colombia Figure 7. Results of crustal thickness estimation beneath 14 of the broad band stations of the NSNC. For the rest of the stations,Receiver Function results were inconclusive. The maximum crustal thickness is about 58 km. Eastern Cordillera Central Cordillera 0 5 10 15 20 25 30 P – S arrival time (s) FLO2 CAP2 BOCO 0 5 10 15 20 25 30 35 40 45 P-arrival time – Origin time (s) 0 5 10 15 20 25 30 P-arrival time – Origin time (s) 0 2 4 6 8 10 12 14 16 18 20 Time (s) 0 2 4 6 8 10 12 14 16 18 20 Time (s)