Times Of Increased Probability Research Articles

Testing an Earthquake Prediction Algorithm: The 2016 New Zealand and Chile Earthquakes

The 13 November 2016, M7.8, 54 km NNE of Amberley, New Zealand and the 25 December 2016, M7.6, 42 km SW of Puerto Quellon, Chile earthquakes happened outside the area of the on-going real-time global testing of the intermediate-term middle-range earthquake prediction algorithm M8, accepted in 1992 for the M7.5+ range. Naturally, over the past two decades, the level of registration of earthquakes worldwide has grown significantly and by now is sufficient for diagnosis of times of increased probability (TIPs) by the M8 algorithm on the entire territory of New Zealand and Southern Chile as far as below 40°S. The mid-2016 update of the M8 predictions determines TIPs in the additional circles of investigation (CIs) where the two earthquakes have happened. Thus, after 50 semiannual updates in the real-time prediction mode, we (1) confirm statistically approved high confidence of the M8–MSc predictions and (2) conclude a possibility of expanding the territory of the Global Test of the algorithms M8 and MSc in an apparently necessary revision of the 1992 settings.

An application of earthquake prediction algorithm M8 in eastern Anatolia at the approach of the 2011 Van earthquake

On 23rd October 2011, an M7.3 earthquake near the Turkish city of Van, killed more than 600 people, injured over 4000, and left about 60,000 homeless. It demolished hundreds of buildings and caused great damages to thousand others in Van, Ercis, Muradiye, and Çaldıran. The earthquake’s epicenter is located about 70 km from a preceding M7.3 earthquake that occurred in November 1976 and destroyed several villages near the Turkey–Iran border and killed thousands of people. This study, by means of retrospective application of the M8 algorithm, checks to see if the 2011 Van earthquake could have been predicted. The algorithm is based on pattern recognition of Times of Increased Probability (TIP) of a target earthquake from the transient seismic sequence at lower magnitude ranges in a Circle of Investigation (CI). Specifically, we applied a modified M8 algorithm adjusted to a rather low level of earthquake detection in the region following three different approaches to determine seismic transients. In the first approach, CI centers are distributed on intersections of morphostructural lineaments recognized as prone to magnitude 7 + earthquakes. In the second approach, centers of CIs are distributed on local extremes of the seismic density distribution, and in the third approach, CI centers were distributed uniformly on the nodes of a 1∘×1∘ grid. According to the results of the M8 algorithm application, the 2011 Van earthquake could have been predicted in any of the three approaches. We noted that it is possible to consider the intersection of TIPs instead of their union to improve the certainty of the prediction results. Our study confirms the applicability of a modified version of the M8 algorithm for predicting earthquakes at the Iranian–Turkish plateau, as well as for mitigation of damages in seismic events in which pattern recognition algorithms may play an important role.

Operational Earthquake Forecasting and Risk Management

In daily life, the general public is commonly exposed to remote uncertain hazards, which would pose a threat to life and property if they were to materialize. Hazard warnings elicit a wide range of precautionary responses, many of which involve comparatively little cost. Air passengers buckle their safety belts during turbulence just in case they may be ejected from their seats. Homeowners in the Caribbean board up their windows with the approach of a tropical storm, just in case it intensifies into a damaging hurricane. U.S. citizens have bought duct tape to seal their homes, just in case of a terrorist attack involving toxic aerosols. Most major earthquakes occur without any observed precursory evidence on which to base a hazard warning decision. However, for some major earthquakes, a degree of precursory evidence may be available that identifies a time of increased probability (TIP). The task of quantifying the probability gain is being made easier by steady technological progress, such as in the higher resolution of fault monitoring. But the level of confidence in an earthquake forecast is extremely unlikely ever to justify the mandatory evacuation of a populous region: the toll of casualties anticipated from the mass evacuation process may be deterrent enough. Notwithstanding this ultimate limitation, a range of far less draconian, and more acceptable, risk mitigation measures may be warranted provided the expected benefits are commensurate with the costs. None of these measures need be mandatory or coercive. Indeed, they may be entirely voluntary. A lesson from the politically influential public policy discourse of Thaler and Sunstein (2008) is that citizens and organizations can be gently “nudged” into taking courses of safer action, without undue authoritarian pressure. Accordingly, the horizon of earthquake forecasting can be expanded far beyond the call for evacuation to justify and motivate practical voluntary …

Predicting the Human Losses Implied by Predictions of Earthquakes: Southern Sumatra and Central Chile

Predictions of earthquakes worldwide by the M8-MSc algorithm, which defines locations of Times of Increased Probability (TIPs), have been tested for nearly two decades, and the authors claim a high rate of success. Thus, it might be appropriate to ask what the consequences in terms of human losses may be if the expected earthquakes should occur. The loss estimating tool QUAKELOSS also has been tested in real-time mode during the last five years with success. Therefore, it is reasonable to estimate the order of magnitude of human losses if great earthquakes should occur in TIPs. Here I compare the consequences if M 8.5 earthquakes should happen in the current TIPs of southern Sumatra and central Chile (Kossobokov and Soloviev, 2008, centers at 4.75S/102.625E and 31.25S/71.77 W, respectively). The selection of the attenuation function is calibrated by matching theoretically calculated intensities and fatalities to the observed values in historic earthquakes. In both areas, the standard attenuation function I use is applicable. The results show that in southern Sumatra fatalities are expected to number fewer than 1,000 (possibly as much as a factor of 5 fewer), whereas they are likely to be larger than 1,000 (possibly as much as a factor six) in central Chile. These figures, however, do not account for possible tsunami effects. The difference is due to two factors. The earthquake sources are farther offshore, and there are only small settlements along the coast in southern Sumatra, whereas along the Chilean coast, large harbor cities are located in the northern part of the TIP area. Regardless of TIP predictions, large earthquakes are to be expected along the Chilean coast. Therefore, it seems advisable to implement mitigating measures in La Serena and Coquimbo, where most of the victims are expected.

Seismic pattern recognition techniques to predict large eruptions at the Popocatépetl, Mexico, volcano

Using pattern recognition techniques, we formulate a simple prediction rule for a retrospective prediction of the three last largest eruptions of the Popocatépetl, Mexico, volcano that occurred on 23 April–30 June 1997 (Eruption 1; VEI ~ 2–3); 11 December 2000–23 January 2001 (Eruption 2; VEI ~ 3–4) and 7 June–4 September 2002 (Eruption 3; explosive dome extrusion and destruction phase). Times of Increased Probability (TIP) were estimated from the seismicity recorded by the local seismic network from 1 January 1995 to 31 December 2005. A TIP is issued when a cluster of seismic events occurs under our algorithm considerations in a temporal window several days (or weeks) prior to large volcanic activity providing sufficient time to organize an effective alert strategy. The best predictions of the three analyzed eruptions were obtained when averaging seismicity rate over a 5-day window with a threshold value of 12 events and declaring an alarm for 45 days. A TIP was issued about six weeks before Eruption 1. TIPs were detected about one and four weeks before Eruptions 2 and 3, respectively. According to our objectives, in all cases, the observed TIPs would have allowed the development of an effective civil protection strategy. Although, under our model considerations the three eruptive events were successfully predicted, one false alarm was also issued by our algorithm. An analysis of the epicentral and depth distribution of the local seismicity used by our prediction rule reveals that successful TIPs were issued from microearthquakes that took place below and towards SE of the crater. On the contrary, the seismicity that issued the observed false alarm was concentrated below the summit of the volcano. We conclude that recording of precursory seismicity below and SE of the crater together with detection of TIPs as described here, could become an important tool to predict future large eruptions at Popocatépetl. Although our model worked well for events that occurred in the past, it is necessary to verify the real capability of the model for future eruptive events.

Diagnosis of Time of Increased Probability (TIP) for Volcanic Earthquakes at Mt. Vesuvius

The possibility of intermediate-term earthquake prediction at Mt. Vesuvius by means of the CN algorithm is explored. CN was originally designed to identify the Times of Increased Probability (TIPs) for the occurrence of strong tectonic earthquakes, with magnitude M ≥ M0, within a region a priori delimited. Here the CN algorithm is applied, for the first time, to the analysis of volcanic seismicity. The earthquakes recorded at Mt. Vesuvius during the period from February 1972 to June 2004 are considered, and the magnitude threshold M0 selecting the events to be predicted is varied within the range: 3.0–3.3. Satisfactory prediction results are obtained, by retrospective analysis, when a time scaling is introduced. In particular, when the length of the time windows is reduced by a factor 2.5–3, with respect to the standard version of CN algorithm, more than 90% of the events with M ≥ M0 occur within the TIP intervals, with TIPs occupying about 30% of the total time considered. The control experiment ``Seismic History'' demonstrates the stability of the obtained results and indicates that the CN algorithm can be applied to monitor the preparation of impending earthquakes with M ≥ 3.0 at Mt. Vesuvius.

Stability of premonitory seismicity pattern and intermediate-term earthquake prediction in Central Italy

The algorithm CN makes use of normalized functions. Therefore the original algorithm, developed for the California-Nevada region, can be directly applied, without adjustment of the parameters, to the determination of the Time of Increased Probability (TIP) of strong earthquakes for Central Italy. The prediction is applied to the events with magnitudeM≥M0=5.6, which in Central Italy have a return period of about six years. The routinely available digital earthquake bulletins of the Istituto Nazionale di Geofisica (ING), Rome, permits continuous monitoring. Here we extend to November 1994 the first study made by Keilis-Boroket al. (1990b). On the basis of the combined analysis of seismicity and seismotectonic, we formulate a new regionalization, which reduces the total alarm time and the failures to predict, and narrows the spatial uncertainty of the prediction with respect to the results ofKeilis-Boroket al. (1990b).

Times of increased probability of large earthquakes ( Ms ≥ 7.5) along the Mexican subduction zone

Using the method of pattern recognition of infrequent events, we explore the possibility of identifying the times of increased probability (TIP) of occurrence of large earthquakes ( M s ≥ 7.5) along the Mexican subduction zone. A TIP refers to a 5 year period within which a strong earthquake has a high probability of occurrence. We analyzed the seismicity for the regions of Chiapas-Eastern Oaxaca (91–97°W, Region 1), Western Oaxaca-Guerrero (97–102°W, Region 2), and Michoacan-Colima-Jalisco (102–106°W, Region 3) for the time interval 1970–1991 with the National Oceanic and Atmospheric Administration seismic catalog as the source data. The premonitory intermediate-term seismic activation to identify a TIP was measured using the algorithm M8. Three out of the five earthquakes of M s ≥ 7.5 that occurred during the studied period are diagnosed by algorithm M8, and for the Region 3 segment a TIP is identified that will end between 1994 and 1996. We consider the area between the Western Colima gap and the Jalisco region (103.7–106.0°W) as the zone where the forecast earthquake has a higher than usual probability of occurrence. Our results by no means constitute a definitive earthquake prediction but suggest the need for detailed seismic analysis and the study of other geophysical precursors along the Colima-Jalisco segment.

A successful prediction of a strong earthquake in the southwestern Hellenic arc by means of the algorithm M8

Abstract In an earlier study, a current TIP (Time of Increased Probability) has been diagnosed along the southwestern part of the Hellenic arc by applying the algorithm M8. In that paper we pointed out that the current TIP of duration of five years started at the end of 1986 or at the beginning of 1987 and if the algorithm M8 works reliably, then it will be completed by the end of 1992 with an earthquake of a surface waves magnitude in the range 6.0 ≤ Ms On November 21, 1992, an earthquake of Ms = 6.5 took place in the candidate region. In the following we discuss the applicability of the algorithm M8 to delineating critical seismic areas and the necessity of using current TIPs to focus the attention of multidisciplinary research for evaluating short-term precursory phenomena, if any.

Times of increased probability of earthquakes of ML ⩾ 5.5 in Greece diagnosed by algorithm M8

An attempt is made to explore the possibility of identifying the Times of Increased Probability (TIPs) of occurrence of earthquakes with M L ⩾ 5.5 by premonitory intermediate-term seismic activation in lower-magnitude ranges, as defined in the algorithm M8. A TIP refers to a 5-year period and an area whose length is several times greater than that of an incipient earthquake. This algorithm has been already tested, retrospectively, for the strongest earthquakes of magnitude greater than or equal to 7.0 which have occurred in Greece since 1973. In the present study, the same algorithm is also applied, retrospectively, by considering smaller earthquakes of local magnitude M L greater than or equal to 5.5. The obtained results show that ten out of eleven events with M ⩾5.5 which took place in Greece and surrounding region from January 1, 1977 till December 31, 1990 occurred within the TIPs diagnosed by the algorithm M 8 retrospectively. Only the Thessaloniki earthquake of June 20, 1978 did not occur within TIP. The M8 algorithm failed in four cases, namely those that occurred in regions for which our catalogue was not complete. Altogether, the TIPs occupy less than 25%, and times expectation about 8% of the total space time considered. There is also a current TIP started at the end of 1986 or at the beginning of 1987 for the south-southwestern part of the Hellenic arc indicating that this area deserves particular attention. If the algorithm M8 works reliably, then this TIP will be filled by the end of 1992. This result by no means constitute the definitive earthquake prediction but merely outline a territory for more detailed and comprehensive analysis.

Diagnosis of time of increased probability of strong earthquakes in the Jordan—Dead Sea rift zone

We have explored the possibility of intermediate-term earthquake prediction in the Jordan-Dead Sea rift region using the CN algorithm (Keilis-Borok et al., 1988). This algorithm is designed to identify the Time of Increased Probability (TIP) of the occurrence of strong earthquakes, defined by the condition M ≥ M 0, where M is the earthquake magnitude and M 0 is a numerical threshold chosen for the considered region. The area under diagnosis includes Israel, most of Lebanon and the eastern parts of Jordan and Syria. Three strong earthquakes defined by the condition M 0 = 5 occurred in this area during the period July 1964-June 1990. By applying the CN algorithm with unchanged standard parameters, we found that two of the strong earthquakes were within the range of corresponding TIP intervals. These intervals extend far beyond the time of the occurrence of the strong earthquakes and represent about 40.5% of the total time. However, the TIP duration is considerably reduced to about 12.9% of the total time, when the threshold of the TIP termination rule is slightly changed. Although this a posteriori change violates the formal standard application of the CN algorithm, the results seem to be encouraging since false alarms did not appear at random, and the excessive TIP duration has a simple and rather obvious explanation. However, it seems that the conditions for the determination of TIPs in the CN algorithm should be modified accordingly.

Intermediate term earthquake prediction in the area of Greece: Application of the algorithm M8

The 3 strongest earthquakes,M≥7.0, which have occurred since 1973 in the area of Greece were preceded by a specific increase of the earthquake activity in the lower magnitude range. This activation is depicted by algorithm M8. This algorithm of intermediate term earthquake prediction was originally designed for diagnosis by Times of Increased Probability (TIPs) of the strongest earthquake,M≥8.0 worldwide (Keilis-Borok andKossobokov, 1984). At present the algorithm is retrospectively tested for smaller magnitudes in different seismic regions (Keilis-Borok andKossobokov, 1986, 1988). A TIP refers to a time period of 5 years and an area whose linear size is proportional and several times larger than that of the incipient earthquake source. Altogether the TIPs diagnosed by the algorithm M8 in the area of Greece occupy less than 20% and the Times of Expectation (TEs) about 10% of the total space-time domain considered. Also there is a current TIP for the southeastern Aegean sea and 1988–1992. It may specify the long-term prediction given inWyss andBaer (1981a,b).

Times of increased probability of strong earthquakes (M ≥ 7.5) diagnosed by Algorithm M8 in Japan and adjacent territories

This paper continues to explore the possibility of identifying the times of increased probability (TIPs) of occurrence of a strong earthquake by premonitory intermediate‐term seismic activation in lower‐magnitude range, as defined in the algorithm M8. A TIP refers to a 5‐year period and an area whose linear size is several times larger than that of an incipient earthquake. In the course of the global test of the algorithm we applied it retrospectively to the data from the National Earthquake Information Center worldwide earthquake catalog (since the regional Japan Meteorological Agency catalog was unavailable to authors), related to the segment of circum‐Pacific belt between latitudes 20°N and 48°N, from the south of Kuril arc to the Taiwan and Izu‐Bonin arc. The worldwide catalog allows us to diagnose TIPs for earthquakes above magnitude 7.5 from 1975 on. Five out of six earthquakes with M ≥ 7.5 occurred within the TIPs diagnosed by the algorithm M8 with a priori fixed parameters. The missed earthquake occurred at the final stage of formation of the corresponding TIP. Altogether, the TIPs occupy less than 20%, and times of expectation occupy about 8%, of the total space‐time domain considered. There are also current TIPs covering about half of the territory. Application of the additional algorithm of localization, named “Mendocino Scenario,” indicates that the southern part of the Japan trench is an area deserving particular attention. The results obtained here by no means constitute the definitive earthquake prediction but merely outline the territories for more detailed and comprehensive analysis. If confirmed by regional data, the results may be used for implementation of some safety measures and evaluation of other precursors, including short‐term ones.

On intermediate-term earthquake prediction in central Italy

The Time of Increased Probability (TIP) for the occurrence of a strong earthquake is determined in Central Italy. This is done with an algorithm that has been successfully applied in other regions of the world (algorithm CN, from the initials of California and Nevada, where the first diagnoses of TIPs were made). The use of normalized functions allows direct application of the orginal algorithm to the new region being studied, without anyad hoc adjustment of the parameters. Retrospective analysis carried on until 1986 shows that TIPs occupy 26 percent of the total time considered and precede four out of five strong earthquakes. Forward monitoring indicates the possible existence of a TIP started in May 1988. Several tests indicate that the results obtained are quite stable, even when using catalogues from different agencies. Apart from obvious practical interest, this research is essential for the worldwide investigation of self-similarity in the origin of strong earthquakes.

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.

Premonitory seismicity patterns near Vancouver Island, Canada

The search for premonitory patterns for large seismic events requires an a-priori threshold to define them. In the area of the Explorer and Juan de Fuca plates (the area of Vancouver Island, Canada) this threshold appears to be M ≥ 6.5. The three strongest earthquakes ( M > 6.0) which have occurred since the seismicity catalog has been acceptably complete for M >- 2.5 are preceded by activation of the earthquake flow in the lower magnitude range, in the generalized sense depicted by the algorithm “M8” suggested by Soviet scientists. One event of magnitude 6.0 may have been preceded by activation, but two others were not. The times of increased probability (TIP) of a strong earthquake occupy about 30% of the total time-space domain considered and there are no “false alarms”. This implies a significant similarity between the earthquake flow here and in other regions of the world where algorithm M8 was tested. There is a possibility, which remains to be tested, of anticipating future strong earthquakes in this area with an accuracy of 5 years and 100 km. Such anticipation could justify some precautionary measures and could be useful for attempts to make short-term predictions. Analysis using patterns of quiescence restrict the location of the anticipated earthquake. There is evidence for the coupling of seismicity in the subduction zone with that above it. Such interaction is a natural consequence of the model of a lithosphere as an assemblage of interacting blocks.