Evaluation of earthquake locations by the two local and regional seismic networks in Central Alborz (Iran)

The reliable location of earthquakes is one of the most important criteria in the assessments of the seismicity, seismotectonics and seismic hazard analysis and focal depth accuracy has been more questionable. In this study, archived data by local temporary seismological network at the southern edge of Central Alborz (Abbassi et al., 2010; Abbassi et al., 2012; Tatar et al., 2012) and by Iranian Seismological Center (IrSC) at the same times are evaluated. The most basic goal of seismic networks is the determination of accurate earthquake locations (Havskov et al, 2012). The local network (TEFIR) has 47 short period stations that are installed at a spacing of 7 to 15 km and worked at 22 weeks in June to November 2006 in area with 35.4° to 36.1° N and 51.5° to 53° E coordinates which operated in continuous and trigger modes at a sampling rate of 100 up to 125 Hz by International Institute of Earthquake Engineering and Seismology (IIEES). In TEFIR network, the time was calibrated every hour by a GPS receiver connected to each station.
The IrSc network is comprised of 17 telemetries, 1-s short-period stations at a spacing of 35 up to 90 km in region with 35° to 37° N and 50° to 54° E coordinates which are operated in continuous mode at a sampling rate of 50 Hz by Institute of Geophysics, University of Tehran (IGUT).
The location improvement for few events in IrSC catalogue, even if being for the limited area and time period, is one of the main goal. Attention to the seismic network, especially station intervals to modify the standards of location criteria is another aim of this attempt.
To do this study, the events reported in S-files format by the both local and regional network catalogues are relocated with Vp/Vs ratio and one dimension crustal velocity structure (Abbassi et al., 2010). This relocation process has been done in SEISAN software (Havskov and Ottemöller, 2005) and Hypocenter program (Lienart et al., 1986; Lienart, 1991; Lienart and Havskov, 1995). Then relevant location parameters for events that are recorded commonly by the two networks at the same times and different only up to 5 seconds, have been compared carefully.
Average error estimations for the same events at all azimuthal gap (Gap) was up to 1.1 seconds in origin times, from 5.3 to 6.9 kilometers for latitudes and longitudes respectively and more than 15 kilometers for focal depths that recorded by the regional network. These estimated errors for the same events in the local network are less than 0.52 second in origin times, 1.9 kilometers for latitudes and longitudes and up to 5.2 kilometers for focal depths (TEFIR201 and IrSC201 in the table 4). Then 27 anticipated well location events have been selected and analyzed based on Gap ≤ 180° simultaneously from the 201data sets in both relocated catalogues (TEFIR27 and IrSC27).
Location criteria for the events in TEFIR27 data set that have proper mean minimum distances to the recorded stations (Dmin) were being more reliable locations. While the expected improve ment in locations for the IrSC27 data set, despite proper Dmin was not satisfactory.
The relocated same events have shown good relative fitness in location dependent parameters such as root mean square residual timing (RMS), Gap, Dmin, number of stations (NST) and number of phases. Also, this proportionality for calculated latitude and longitude coordinates and origin times, is more than for the calculated depths.
This approaches show that the uncertainties have specific dependencies on the input data files, especially phase readings and phase pickings of the events. As were observed in some seismic records by the regional network at the stations and the reread of the Pg and Sg phases by the local network, some tendency to RMS reduction in the regional operation alone, and intern phases of the far stations to the event locations scheme were being some sources of location errors. Reading the Pn phase instead of the Pg at under the critical distances to receiver station for example less than 180 kilometers, and phase detecting error for few reported events were most erronious cases.
Obviously, reliable crustal model and Vp/Vs ratio in addition to other appropriate location parameters with a simultaneous decrease RMS causes better location accuracies. Then application of the proper criteria such as Dmin, Gap, NST and corrected phase pickings may improve related location parameters, containing RMS, spatial and time accuracies under the other restrictive network conditions. Among these criteria, Dmin has an important role in determination of the seismological network geometries.
The results of this study, despite the fundamental difference in the processing methods and data with the work of Bond´ar et al. (2004), have a good agreement in the discussed criteria and analytical issues related to reliable locations. Also the consistency of this work may be useful for confidence of the location certainties at least for some earthquake occurred in the studied area at the time span.