Magnitude-scaling relations using period parameters τc and τpmax, for Tehran region, Iran

Abstract

In this study, the first step towards establishing an onsite earthquake early warning system (EWS) in the Tehran region is presented. The system uses the period parameters tau(max)(p) and tau(c) from the first 3 s of the vertical and horizontal components of a P wave, separately and combined. Various regression relations between the magnitude and period parameters were determined for different seismic networks operating in the study area. The data set used in this study contains small ground motions including 194 events with magnitudes between M-L 2.5 and 4.6 located within approximately 80 km from the epicentre in the Tehran region. The SDs of the magnitude-scaling relations for all the component categories (vertical, horizontal and total components) based on the tau(max)(p) and tau(c) approaches were estimated to be on the order of +/- 1.0 and +/- 1.1 unit of magnitude, respectively. These relations were determined from the small magnitude range of the velocity records (M-L 2.5-4.6) as input seismograms. Additional tests were conducted to verify the reliability and robustness of the determined magnitude-scaling relations using acceleration records from the 2002 June 22, M-L 6.5 Changureh-Avaj earthquake; 2004 May 28, M-L 6.1 Firoozabad-Koojour earthquake; 2009 October 17, M-L 3.9 Shahre-Rey earthquake and 2011 February 20, M-L 4.1 Sharif-Abad earthquake; the first two events (Changureh-Avaj and Firoozabad-Koojour) occurred outside the study area. Among the various regression scaling relations obtained, the estimated magnitude based on the tau(max)(p) approach using the vertical components yielded the most stable and reliable results of 6.4 (+/- 0.4), 5.9 (+/- 0.4), 3.3 (+/- 0.5) and 3.8 (+/- 0.3) for the Changureh-Avaj, Firoozabad-Koojour, Shahre-Rey and Sharif-Abad earthquakes, respectively. The magnitudes estimated using the tau(c) method exhibited more scatter with higher SDs than those using the vertical components using the tau(max)(p) approach. Our results also indicate that using the horizontal components produces larger SDs, which may be attributed to the larger site effects; however, the horizontal components can be used as auxiliary available data to provide more constrained information for a multilevels pilot alarm system and to reduce the number of missed or false alarms. The main uncertainties in the proposed magnitude-scaling relations result from the absence of any large earthquakes and poor station distributions in the study area. However, the results presented in this study can be used as a pilot onsite earthquake EWS in the Tehran region.