Vladimir I. Keilis-borok

Premonitory activation of earthquake flow : algorithm M8

Abstract Thirty-nine out of the 44 strongest earthquakes which have recently occurred in different regions of the world are preceded by specific activation of the earthquake flow in the lower magnitude range. This activation is depicted by the algorithm M8, which was designed for diagnosis of times of increased probability (TIPs) of strong earthquakes. A TIP refers to a time period of 5 yr and an area of linear size several times larger than that of the incipient earthquake source. Altogether the TIPs diagnosed in this study occupy less than 20%, and the times of expectation (TEs) about 10% of the total space-time domain considered. These results constitute an argument in favor of at least partial self-similarity of the earthquake flow in diverse seismotectonic environments and magnitude ranges. It also implies the possibility of intermediate-term prediction of the strongest earthquakes in these regions. Such prediction may be used for precautionary measures.

Nonlinear dynamics of the lithosphere and earthquake prediction

1 Fundamentals of Earthquake Prediction: Four Paradigms.- 2 Hierarchical Models of Seismicity.- 3 Models of Dynamics of Block-and-Fault Systems.- 4 Earthquake Prediction.- 5 Earthquake Prediction Strategies: A Theoretical Analysis.- 6 Recognition of Earthquake-Prone Areas.- References.

Diagnosis of Time of Increased Probability of strong earthquakes in different regions of the world: algorithm CN

Abstract An algorithm for intermediate-term earthquake prediction is suggested which allows diagnosis of the times of increased probability of strong earthquakes (TIPs). TIPs are declared for the time period of one year and an area with linear dimensions of a few hundred kilometers, and can be extended in time. The algorithm is based on the following traits of an earthquake flow: level of seismic activity; its temporal variation; clustering of earthquakes in space and time; their concentration in space; and their long-range interaction. The algorithm is normalized so that it can be applied in various regions without readaptation. TIPs, diagnosed by the algorithm, precede ∼ 80% of strong earthquakes and take on average ∼ 24% of the total time.

Clustering analysis of seismicity and aftershock identification

We introduce a statistical methodology for clustering analysis of seismicity in the time-space-energy domain and use it to establish the existence of two statistically distinct populations of earthquakes: clustered and nonclustered. This result can be used, in particular, for nonparametric aftershock identification. The proposed approach expands the analysis of Baiesi and Paczuski [Phys. Rev. E 69, 066106 (2004)10.1103/PhysRevE.69.066106] based on the space-time-magnitude nearest-neighbor distance eta between earthquakes. We show that for a homogeneous Poisson marked point field with exponential marks, the distance eta has the Weibull distribution, which bridges our results with classical correlation analysis for point fields. The joint 2D distribution of spatial and temporal components of eta is used to identify the clustered part of a point field. The proposed technique is applied to several seismicity models and to the observed seismicity of southern California.

Testing earthquake prediction algorithms: statistically significant advance prediction of the largest earthquakes in the Circum-Pacific, 1992–1997

Abstract Algorithms M8 and MSc (i.e., the Mendocino Scenario) were used in a real-time intermediate-term research prediction of the strongest earthquakes in the Circum-Pacific seismic belt. Predictions are made by M8 first. Then, the areas of alarm are reduced by MSc at the cost that some earthquakes are missed in the second approximation of prediction. In 1992–1997, five earthquakes of magnitude 8 and above occurred in the test area: all of them were predicted by M8 and MSc identified correctly the locations of four of them. The space–time volume of the alarms is 36% and 18%, correspondingly, when estimated with a normalized product measure of empirical distribution of epicenters and uniform time. The statistical significance of the achieved results is beyond 99% both for M8 and MSc. For magnitude 7.5+, 10 out of 19 earthquakes were predicted by M8 in 40% and five were predicted by M8–MSc in 13% of the total volume considered. This implies a significance level of 81% for M8 and 92% for M8–MSc. The lower significance levels might result from a global change in seismic regime in 1993–1996, when the rate of the largest events has doubled and all of them become exclusively normal or reversed faults. The predictions are fully reproducible; the algorithms M8 and MSc in complete formal definitions were published before we started our experiment [Keilis-Borok, V.I., Kossobokov, V.G., 1990. Premonitory activation of seismic flow: Algorithm M8, Phys. Earth and Planet. Inter. 61, 73–83; Kossobokov, V.G., Keilis-Borok, V.I., Smith, S.W., 1990. Localization of intermediate-term earthquake prediction, J. Geophys. Res., 95, 19763–19772; Healy, J.H., Kossobokov, V.G., Dewey, J.W., 1992. A test to evaluate the earthquake prediction algorithm, M8. U.S. Geol. Surv. OFR 92-401]. M8 is available from the IASPEI Software Library [Healy, J.H., Keilis-Borok, V.I., Lee, W.H.K. (Eds.), 1997. Algorithms for Earthquake Statistics and Prediction, Vol. 6. IASPEI Software Library].

Reverse tracing of short-term earthquake precursors

Abstract We introduce a new approach to short-term earthquake prediction named “ Reverse Tracing of Precursors ” (RTP), since it considers precursors in reverse order of their appearance. First, we detect the “candidates” for the short-term precursors; in our case, these are newly introduced chains of earthquakes reflecting the rise of an earthquake correlation range. Then we consider each chain, one by one, checking whether it was preceded by an intermediate-term precursor in its vicinity. If yes , we regard this chain as a precursor; in prediction it would start a short-term alarm. The chain indicates the narrow area of possibly complex shape, where an intermediate-term precursor should be looked for. This makes possible to detect precursors undetectable by the direct analysis. RTP can best be described on an example of its application; we describe retrospective prediction of two prominent Californian earthquakes—Landers (1992), M =7.6, and Hector Mine (1999), M =7.3, and suggest a hypothetical prediction algorithm. This paper descripes the RTP methodology, which has potentially important applications to many other data and to prediction of other critical phenomena besides earthquakes. In particular, it might vindicate some short-term precursors, previously rejected as giving too many false alarms. Validation of the algorithm per se requires its application in different regions with a substantial number of strong earthquakes. First (and positive) retrospective results are obtained for 21 more strong earthquakes in California ( M ≥6.4), Japan ( M ≥7.0) and the Eastern Mediterranean ( M ≥6.5); these results are described elsewhere. The final validation requires, as always, prediction in advance for which this study sets up a base. We have the first case of a precursory chain reported in advance of a subsequent strong earthquake (Tokachi-oki, Japan, 25 September 2003, M =8.1). Possible mechanisms underlying RTP are outlined.

A Boolean Delay Equation Model of Colliding Cascades. Part II: Prediction of Critical Transitions

We consider here prediction of abrupt overall changes (“critical transitions”) in the behavior of hierarchical complex systems, using the model developed in the first part of this study. The model merges the physical concept of colliding cascades with the mathematical framework of Boolean delay equations. It describes critical transitions that are due to the interaction between direct cascades of loading and inverse cascades of failures in a hierarchical system. This interaction is controlled by distinct delays between switching of elements from one state to another: loaded vs. unloaded and intact vs. failed. We focus on the earthquake prediction problem; accordingly, the models heuristic constraints are taken from the dynamics of seismicity. The model exhibits four major types of premonitory seismicity patterns (PSPs), which have been previously identified in seismic observations: (i) rise of earthquake clustering; (ii) rise of the earthquakes intensity; (iii) rise of the earthquake correlation range; and (iv) certain changes in the size distribution of earthquakes (Gutenberg–Richter relation). The model exhibits new features of individual PSPs and their collective behavior, to be tested in turn on observations. There are indications that the premonitory phenomena considered are not seismicity-specific, but may be common to hierarchical systems of a more general nature.

PREMONITORY TRANSFORMATION OF STEEL FRACTURING AND SEISMICITY

Abstract Transformation of microfracturing preceding the break up of steel and rock samples is similar to transformation of earthquakes flow prior to strong earthquakes in Southern California. The break up in a sample is preceded by transition from formation of new microcracks to coalescence or expansion of existing cracks. This is reflected in the relation between the length and the number of the cracks. Similar transformation of seismicity precedes Southern California earthquakes with magnitude M ≥ 6.6, 1935–1994. Specifically, magnitude-frequency relation bends downward for magnitudes from 3 to 4.5 and upward for magnitudes from about 4.5 to 6; in the usual notations the ‘b-value’ becomes larger in the first interval and smaller in the second one. This transformation is accompanied by the increased share of aftershocks in the earthquakes flow. In such a way the approach of a strong earthquake is reflected in both major traits of seismicity: magnitude-frequency relation and earthquake clustering. Imprecision of the earthquake catalog and reasonable variations in its analysis do not change our conclusions. This phenomenon explains, so far—qualitatively, a wide set of premonitory seismicity patterns. We give it a formal definition, allowing to test whether it takes place in other seismic regions.

Earthquake prediction: Probabilistic aspect

SUMMARY A theoretical analysis of the earthquake prediction problem in space‐time is presented. We find an explicit structure of the optimal strategy and its relation to the generalized error diagram. This study is an extension of the theoretical results for time prediction. The possibility and simplicity of this extension is due to the choice of the class of goal functions. The generalized error diagram allows us to suggest a natural measure of prediction efficiency at the research stage.

Pre‐recession pattern of six economic indicators in the USA

This paper applies a tightly parameterized pattern recognition algorithm, previously applied to earthquake prediction, to the problem of predicting recessions. Monthly data from 1962 to 1996 on six leading and coincident economic indicators for the USA are used. In the full sample, the model performs better than benchmark linear and non-linear models with the same number of parameters. Subsample and recursive analysis indicates that the algorithm is stable and produces reasonably accurate forecasts even when estimated using a small number of recessions. Copyright