Maria Lancieri

Earthquake magnitude estimation from peak amplitudes of very early seismic signals on strong motion records

[1] We show that the low-pass filtered, peak amplitudes of initial P- and S-wave seismic signals recorded in the vicinity of an occurring earthquake source correlates with the earthquake magnitude and may be used for real-time estimation of the event size in seismic early warning applications. The earthquake size can be therefore estimated using only a couple of seconds of signal from the P- or S-wave onsets, i.e. while the rupture itself is still propagating and rupture dimension is far from complete. We argue that dynamic stress release and/or slip duration on the fault in the very early stage of seismic fracture, scales both with the observed peak amplitude and with the elastic energy available for fracture propagation. The probability that a fracture grows to a larger size should scale with the energy initially available. Citation: Zollo, A., M. Lancieri, and S. Nielsen (2006), Earthquake magnitude estimation from peak amplitudes of very early seismic signals on strong motion records, Geophys. Res. Lett., 33, L23312, doi:10.1029/ 2006GL027795.

Earthquake early warning system in southern Italy: Methodologies and performance evaluation

We investigate the effect of extended faulting processes and heterogeneous wave propagation on the early warning system capability to predict the peak ground velocity (PGV) from moderate to large earthquakes occurring in the southern Apennines (Italy). Simulated time histories at the early warning network have been used to retrieve early estimates of source parameters and to predict the PGV, following an evolutionary, probabilistic approach. The system performance is measured through the Effective Lead-Time (ELT), i.e., the time interval between the arrival of the first S-wave and the time at which the probability to observe the true PGV value within one standard deviation becomes stationary, and the Probability of Prediction Error (PPE), which provides a measure of PGV prediction error. The regional maps of ELT and PPE show a significant variability around the fault up to large distances, thus indicating that the systems capability to accurately predict the observed peak ground motion strongly depends on distance and azimuth from the fault.

Reply to comment by P. Rydelek et al. on ‘‘Earthquake magnitude estimation from peak amplitudes of very early seismic signals on strong motion records’’

[1] Based on the analysis of Mediterranean, near-source, strong motion records Zollo et al. [2006] (hereinafter referred to as ZLN) showed that peak displacement amplitudes of initial Pand S-wave seismic signals scales with the earthquake size in the moment magnitude range 4 6.1. [9] In Figure 1 we report data distribution vs. magnitude and hypocentral distance. Note that the distance distribution of analyzed events for Japan (Figure 1) is clearly different from that for the Mediterranean (MED) earthquakes given by ZLN. Whereas most of MED records occur at an HD 40 km. [10] At odds with RWH, we did not apply any cut-off at M = 5.5 on hypocentral distance for small and moderate events; indeed this may introduce a bias on the estimation of attenuation law parameters retrieved from both small and large earthquakes records. We follow the ZLN procedure involving among others sensitivity correction, P and S phase identification, double integration, filtering in a frequency band (0.075, 3 Hz). Peaks are read on the modulus of displacement expressed in meters. Starting from the Pwave and S-wave picked arrivals, we measure peaks within a 2 second window. For the P arrival, we also test the method with 4 second windows. [11] The distance attenuation effect is corrected for by assuming a simple log-linear model [Wu and Zhao, 2006; ZLN]; the amplitudes are normalized to a reference distance of 10 Km (i.e., the same as for ZLN in the Mediterranean study, to facilitate comparison). Though RWH show data normalized at a distance of 20 km (ZLN, Figures 2a and 2b), the use of a different normalization distance does not change the slope of the log(peak) vs Magnitude relationships and thus does not alter the graphical interpretation. [12] In Figure 2 we show the relation between the logarithm of distance normalized, peak displacement for P and S waves in different time windows vs final magnitude of the events (following ZLN, statistical significance was GEOPHYSICAL RESEARCH LETTERS, VOL. 34, L20303, doi:10.1029/2007GL030560, 2007 Click Here for Full Article

Short-Period Rupture Process of the 2010 Mw 8.8 Maule Earthquake in Chile

The 2010 Maule earthquake is one of the largest events ever recorded with modern instruments. We used the continuous GPS (cGPS) records to invert for the kinematic rupture process using an elliptical sub-patch approximation. In agreement with previous inversions, the largest slip is found in the northern part of the rupture zone. By cross-correlating signals from cGPS and strong motion records (SM) located in the northern part of the rupture zone, we identified two distinct seismic pulses. Using the arrival time of these pulses, we propose a short-period (<20 s) rupture process, the zone where these pulses are generated is situated near 35.5°S, in agreement with the area with the highest seismic slip and maximum observed intensity. Finally, we compare the strong motion records at the same sites for the 1985 Mw 8 Valparaíso earthquake and the Maule earthquake. We found that spectral contents and duration of the records of these two events were very similar. Thus, at least in the northern part of the rupture, the Maule earthquake radiated high frequency waves like an Mw 8 earthquake.

An evolutionary approach to real‐time moment magnitude estimation via inversion of displacement spectra

[1] We present an evolutionary approach for magnitude estimation for earthquake early warning based on real‐time inversion of displacement spectra. The Spectrum Inversion (SI) method estimates magnitude and its uncertainty by inferring the shape of the entire displacement spectral curve based on the part of the spectra constrained by available data. The method consists of two components: 1) estimating seismic moment by finding the low frequency plateau W0, the corner frequency fc and attenuation factor (Q) that best fit the observed displacement spectra assuming a Brune w 2 model, and 2) estimating magnitude and its uncertainty based on the estimate of seismic moment. A novel characteristic of this method is that is does not rely on empirically derived relationships, but rather involves direct estimation of quantities related to the moment magnitude. SI magnitude and uncertainty estimates are updated each second following the initial P detection. We tested the SI approach on broadband and strong motion waveforms data from 158 Southern California events, and 25 Japanese events for a combined magnitude range of 3 ≤ M ≤ 7. Based on the performance evaluated on this dataset, the SI approach can potentially provide stable estimates of magnitude within 10 seconds from the initial earthquake detection. Citation: Caprio,M.,M.Lancieri,G.B.Cua,A.Zollo, and S. Wiemer (2011), An evolutionary approach to real‐time moment magnitude estimation via inversion of displacement spectra, Geophys. Res. Lett., 38, L02301, doi:10.1029/ 2010GL045403.

Real-Time Estimation of Earthquake Magnitude for Seismic Early Warning

A prototype system for earthquake early warning and rapid shake map evaluation is being developed and tested in southern Italy based on a dense, dynamic seismic network (accelerometers + seismometers) under installation in the Apenninic belt region (Irpinia Seismic Network). It can be classified as a regional Earthquake Early Warning System consisting of a broad-based seismic sensor network covering a portion or the entire area which is threatened by the quake’s strike.