Earthquake pattern analysis using subsequence time series clustering

In this paper, a subsequence time-series clustering algorithm is proposed to identify the strongly coupled aftershocks sequences and Poissonian background activity from earthquake catalogs of active regions. The proposed method considers the inter-event time statistics between the successive pair of events for characterizing the nature of temporal sequences and observing their relevance with earthquake epicenters and magnitude information simultaneously. This approach categorizes the long-earthquake time series into the finite meaningful temporal sequences and then applies the clustering mechanism to the selective sequences. The proposed approach is built on two phases: (1) a Gaussian kernel-based density estimation for finding the optimal subsequence of given earthquake time-series, and (2) inter-event time ( Δ t ) and distance-based observation of each subsequence for checking the presence of highly correlated aftershock sequences (hot-spots) in it. The existence of aftershocks is determined based on the coefficient of variation (COV). A sliding temporal window on Δ t with earthquake's magnitude M is applied on the selective subsequence to filter out the presence of time-correlated events and make the meaningful time stationary Poissonian subsequences. This proposed approach is applied to the regional Sumatra-Andaman (2000-2021) and worldwide ISC-GEM (2000-2016) earthquake catalog. Simulation results indicate that meaningful subsequences (background events) can be modeled by a homogeneous Poisson process after achieving a linear cumulative rate and time-independent λ in the exponential distribution of Δ t . The relations C O V a ( T ) > C O V o ( T ) > ( C O V b ( T ) ≈ 1 ) and C O V a ( d ) > C O V o ( d ) > C O V b ( d ) are achieved for both studied catalogs. Comparative analysis justifies the competitive performance of the proposed approach to the state-of-art approaches and recently introduced methods.

An Earthquake-Clustering Model in North Aegean Area (Greece)

The investigation of short-term earthquake-clustering features is made feasible through the application of a purely stochastic Epidemic-Type Aftershock Sequence (ETAS) model. The learning period that is used for the estimation of the parameters is composed by earthquakes with M ≥ 2.6 that occurred between January 2008 and May 2017. The model predictability is retrospectively examined for the 12 June 2017 Lesvos earthquake (Mw6.4) and the subsequent events. The construction of time-dependent seismicity maps and comparison between the observed and expected earthquake number are performed in order to temporally and spatially test the evolution of the sequence, respectively. The generation of 127 target events with M ≥ 3.0 in the period June–July 2017, just before the main shock occurrence, is examined in a quantitative evaluation. The statistical criteria used for assessing the model performance are the Relative Operating Characteristic Diagram, the R-score, and the probability gain. Reliable forecasts are provided through the epidemic model testifying its superiority towards a time-invariant Poisson model.

Estimation, diagnostics, and extensions of nonparametric Hawkes processes with kernel functions

The Hawkes self-exciting model has become one of the most popular point-process models in many research areas in the natural and social sciences because of its capacity for investigating the clustering effect and positive interactions among individual events/particles. This article discusses a general nonparametric framework for the estimation, extensions, and post-estimation diagnostics of Hawkes models, in which we use the kernel functions as the basic smoothing tool.

Stability of earthquake clustering models: criticality and branching ratios

We study the stability conditions of a class of branching processes prominent in the analysis and modeling of seismicity. This class includes the epidemic-type aftershock sequence (ETAS) model as a special case, but more generally comprises models in which the magnitude distribution of direct offspring depends on the magnitude of the progenitor, such as the branching aftershock sequence (BASS) model and another recently proposed branching model based on a dynamic scaling hypothesis. These stability conditions are closely related to the concepts of the criticality parameter and the branching ratio. The criticality parameter summarizes the asymptotic behavior of the population after sufficiently many generations, determined by the maximum eigenvalue of the transition equations. The branching ratio is defined by the proportion of triggered events in all the events. Based on the results for the generalized case, we show that the branching ratio of the ETAS model is identical to its criticality parameter because its magnitude density is separable from the full intensity. More generally, however, these two values differ and thus place separate conditions on model stability. As an illustration of the difference and of the importance of the stability conditions, we employ a version of the BASS model, reformulated to ensure the possibility of stationarity. In addition, we analyze the magnitude distributions of successive generations of the BASS model via analytical and numerical methods, and find that the compound density differs substantially from a Gutenberg-Richter distribution, unless the process is essentially subcritical (branching ratio less than 1) or the magnitude dependence between the parent event and the direct offspring is weak.

Properties of the probability distribution associated with the largest event in an earthquake cluster and their implications to foreshocks

The space-time epidemic-type aftershock sequence model is a stochastic branching process in which earthquake activity is classified into background and clustering components and each earthquake triggers other earthquakes independently according to certain rules. This paper gives the probability distributions associated with the largest event in a cluster and their properties for all three cases when the process is subcritical, critical, and supercritical. One of the direct uses of these probability distributions is to evaluate the probability of an earthquake to be a foreshock, and magnitude distributions of foreshocks and nonforeshock earthquakes. To verify these theoretical results, the Japan Meteorological Agency earthquake catalog is analyzed. The proportion of events that have 1 or more larger descendants in total events is found to be as high as about 15%. When the differences between background events and triggered event in the behavior of triggering children are considered, a background event has a probability about 8% to be a foreshock. This probability decreases when the magnitude of the background event increases. These results, obtained from a complicated clustering model, where the characteristics of background events and triggered events are different, are consistent with the results obtained in [Ogata, Geophys. J. Int. 127, 17 (1996)] by using the conventional single-linked cluster declustering method.