The Overlap of Aftershock Coda Waves and Short‐Term Postseismic Forecasting

Abstract

The elaboration of reliable forecasting in the first hours after large shocks, very useful for the postseismic management, is strongly affected by the huge incompleteness of seismic catalogs. The deficit of observed events, in the first part of aftershock sequences, can be naturally attributed to different mechanisms such as the inefficiency of the seismic network as well as the overlap of seismic records. In this study we show that short‐term aftershock incompleteness can be explained only in terms of the second mechanism whereas it is only weakly affected by the quality of the instrumental coverage. Indeed, we show that larger deviations from the Gutenberg‐Richter law and the Omori‐Utsu law can be found in regions with a higher seismic network efficiency. We use these results to develop an incomplete Epidemic‐Type Aftershock Sequence model able to reproduce statistical features in time and magnitude of recorded aftershocks.

Post Seismic Catalog Incompleteness and Aftershock Forecasting

“…The same conclusion can be also obtained from the measurement of the correlation between magnitude according to the method proposed in [19,[36][37][38]. This analysis [19,31] has shown significantly larger magnitude correlations in Region 1 than in Region 2.…”

“…The results ( Figure 6) show that the aftershock rate clearly depends on the magnitude difference m M − m th in both Region 1 and Region 2. In particular, de Arcangelis et al [31] divided time by τ = 10 d(m M −m th ) obtaining that data for different values of m M and m th , inside each sub-region, exhibits the scaling collapse ρ(t, m M , m th ) = F(t/τ) ( Figure 7a). It is evident from Figure 7a that the Omori decay ρ ∼ t −p sets in when t/τ becomes larger than a given value x 0 , different between the two regions.…”

“…As explained in Section 2 the static m c is a local quantity which depends on the local density of the seismic network ρ S , as illustrated by Equation (1) and Figure 1. The influence of the density ρ S on the STAI was addressed by de Arcangelis et al [31] by investigating the c meas -value in the two sub-regions of Southern California illustrated in Figure 1. As already explained in Section 2, the inner region (Region 1) comprises a high value of ρ S and a static m c ≤ 1.4.…”

“…Conversely, a small ρ S is present in the external region (Region 2) and the static m c > 1.5, with values of m c 3 close to the borders. To obtain an estimate of c meas in each sub-region, de Arcangelis et al [31] measured the aftershock daily rate ρ(t, m M , m th ) defined as the number of aftershocks with magnitude larger than m th occurring at a temporal distance t after their triggering main shock with magnitude m ∈ [m M , m M + 1), divided by the number of mainshocks with magnitude m ∈ [m M , m M + 1). Three different values of m M = (3, 4, 5) and m th = (1.5, 2.5, 3.5) were considered.…”

“…At the first step, standard ETAS catalogs are simulated and, at the second step, all events that occurred at a temporal distance smaller than ∆t after a larger event are removed from the catalog. de Arcangelis et al [31] implemented different values of K 0 and analyzed the ETASI1 catalog by the same BP declustering procedure applied to the instrumental catalog. As in Figure 6, the aftershock daily rate ρ(t, m M , m th ) for the ETASI1 catalog has been evaluated for different mainshock magnitudes m M , different thresholds m th and different K 0 values.…”