Hydrodynamic and seismic response to teleseismic waves of strong remote earthquakes in Caucasus

Chélidzé, T.; Melikadze, Giorgi; Kobzev, Gennady; Shengelia, Ia; Jorjiashvili, Nato; Mepharidze, Ekaterine

doi:10.1007/s11600-018-00241-7

Cited by 6 publications

(6 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…fast squirt-flow of pore water from the future EQ source to the well [Dvorkin et al, 1994]. Such signals can travel a long distances [Chelidze et al, 2019]. We accept that the interaction length R for geomagnetic precursors of the order of 300 km around Dusheti observatory.…”

Section: The Network Of Observations Preparation Of Data Bases and Me...mentioning

confidence: 77%

Earthquake forecast by imbalance machine learning using geophysical predictors

Kiria,

Chelidze,

Melikadze

et al. 2023

Annals of Geophysics

View full text Add to dashboard Cite

In the present paper we consider the earthquake forecast as a binary problem of machine learning on the imbalanced data base applied to five regions of Georgia. For the training we used geophysical data base collected in 2017-2021, namely, variations of statistical characteristics of geomagnetic field components, seismic activity, water level in deep boreholes and tides. In this version a new predictor – the weighted seismic activity for previous 5 days - – is added compared to the predictors’ list used in previous papers. Besides, the length of the used database is increased 3 times compared to the earlier results. As in the database the earthquakes of M > 3.5 are rare, the number of negative cases is large (there are many days without EQs of M > 3.5), meaning that there is a strong imbalance between positive and negative cases of the order of 1:20; we apply the specific methodology Matthews’ correlation coefficient (MCC) and F1 score to avoid the strong imbalance effect.

show abstract

Section: The Network Of Observations Preparation Of Data Bases and Me...mentioning

confidence: 77%

Earthquake forecast by imbalance machine learning using geophysical predictors

Kiria,

Chelidze,

Melikadze

et al. 2023

Annals of Geophysics

View full text Add to dashboard Cite

show abstract

“…This approach actually does not restrict the interaction length and the delay range for the inducing event of the hydraulic precursor. We presume that there are at least two physical mechanisms, which can explain the accepted long radius of action of hydraulic precursors (R = 200 km): 1) long-range fluid diffusion from the stressed formation to the well due to poroelastic effects and 2) fast squirt-flow (Dvorkin et al, 1994) of pore water from the impending EQ source to the well, excited by the foreshocks of the imminent event-such signals can travel a long distance (Chelidze et al, 2019).…”

Section: Network Devices and Datamentioning

confidence: 99%

Earthquake Forecast as a Machine Learning Problem for Imbalanced Datasets: Example of Georgia, Caucasus

et al. 2022

Self Cite

View full text Add to dashboard Cite

In this article, we considered the problem of M≥3 earthquake (EQ) forecasting (hindcasting) using a machine learning (ML) approach, using experimental (training) time series on monitoring water-level variations in deep wells as well as geomagnetic and tidal time series in Georgia (Caucasus). For such magnitudes’, the number of “seismic” to “aseismic” days in Georgia is approximately 1:5 and the dataset is close to the balanced one. However, the problem of forecast is practically important for stronger events—say, events of M≥3.5—which means that the learning dataset of Georgia became more imbalanced: the ratio of seismic to aseismic days for in Georgia reaches the values of the order of 1:20 and more. In this case, some accepted ML classification measures, such as accuracy leads to wrong predictions due to a large number of true negative cases. As a result, the minority class, here—seismically active periods—is ignored at all. We applied specific measures to avoid the imbalance effect and exclude the overfitting possibility. After regularization (balancing) of the training data, we build the confusion matrix and performed receiver operating classification in order to forecast the next day probability of M≥3.5 earthquake occurrence. We found that the Matthews’ correlation coefficient (MCC) is the measure, which gives good results even if the negative and positive classes are of very different sizes. Application of MCC to observed geophysical data gives a good forecast of the next day M≥3.5 seismic event probability of the order of 0.8. After randomization of EQ dates in the training dataset, the Matthews’ coefficient efficiency decreases to 0.17.

show abstract

“…Scientists do temporal and Spatial Geophysical Data Analysis for the Issues of Natural Hazards and Risk Assessment [Matcharashviliet al, 2015;Melkov et al, 2022]. Scientists study correlation between the value of macroseismic intensity and the indicators of instrumental records [Chelidze et al, 2019;Zaalishvili et al, 2014aZaalishvili et al, , b, 2016Zaalishvili et al, , 2022 and also process some new data on the influence of various soil conditions on probabilistic seismic hazard assessment of territories [Chernov et al, 2022]. In the course of the study, the works of a number of authors were studied: [Yaitskaya, Brigida, 2022…”

Section: Introductionmentioning

confidence: 99%

Анализ Моделей Прогнозирования Землетрясений Для Получения Наилучшей Модели

Pkhovelishvili¹,

Archvadze²,

Nikoleishvili³

2023

Геология И Геофизика Юга России

View full text Add to dashboard Cite

The science of studying the causes of earthquakes is rapidly developing. Each cause of an earthquake can be considered a precursor of an earthquake and using these precursors, predictive models can be built. To date, there are quite a few earthquake prediction models, which allow you to analyze these models to improve accuracy, that is, apply forecast data with higher probabilities. Aim. Analyze forecasting models and, based on the substantiation of existing forecasts, classify them into “necessary” and “sufficient” models, and define these terms. And also, to determine the algorithms for planning further actions to obtain much better forecasting models. It is “necessary” to develop algorithms that bring the “necessary” model to the “sufficient” one and vice versa. “necessary” forecasting models are models whose set of forecasts always includes a set of actually occurring events, and “sufficient” forecasting models are models whose forecasts always come true. The research methodology is to process the existing large data structures that are specified for further use in our algorithm. To calculate the probability of forecast accuracy, an algorithm with "parallel data" - "parallel probability" is used, which allows you to select those pairs of forecasting models (or triples, quadruples, etc.), whose "joint" probability of forecast accuracy gives a much better result than separately. Results were the formation of an author's approach to processing earthquake forecast models and obtaining a generalized model that gives forecasts with a higher probability due to the use of statistics from already existing forecast models and their further observation. Algorithms have been defined for a) when to analyze all available required models and obtain one best model by combining the appropriate number of required models and b) when to combine enough models closest to guessing all predictions so that their number is less than in other unions. Also exists an algorithm that determines the study to be carried out after the occurrence of each event - the calculation of the probability of justification for individual models, as well as paired and triple models. The problem of using these algorithms in a specific area - earthquake prediction is discussed, and the results of the work of the author's algorithm are shown. Наука, изучающая причины землетрясений, стремительно развивается. Каждую причину землетрясения можно рассматривать как предвестник землетрясения, и, используя эти предвестники, построить модели прогнозирования. На сегодняшний день существует довольно много моделей прогнозирования землетрясений, что позволяет анализировать эти модели для повышения точности, то есть применять данные прогноза с более высокими вероятностями. Целью работы является анализ модели прогнозирования, их классификация на основе существующих прогнозов на «необходимые» и «достаточные» модели, дать определение этим понятиям. Также целью является определение алгоритмов планирования дальнейших действий для получения гораздо более качественных моделей прогнозирования. Необходимо разработать алгоритмы, приводящие «необходимую» модель к «достаточной» и наоборот. «Необходимые» модели прогнозирования – это модели, набор прогнозов которых всегда включает набор реально происходящих событий, а «достаточные» модели прогнозирования – это модели, прогнозы которых всегда сбываются. Метод исследования заключается в обработке имеющихся больших структур данных, которые заданы для дальнейшего использования в нашем алгоритме. Для расчета вероятности точности прогноза используется алгоритм с «параллельными данными» – «параллельная вероятность», который позволяет выбрать те пары моделей прогнозирования (или тройки, четверки и т.д.), «совместная» вероятность точности прогноза которых дает гораздо лучший результат, чем по отдельности. Результатами исследования стали формирование авторского подхода к обработке моделей прогноза землетрясений и получение обобщенной модели, дающей прогнозы с большей вероятностью за счет применения статистики уже существующих моделей прогноза и их дальнейшего наблюдения. Алгоритмы были определены для вариантов: а) анализа всех имеющихся необходимых моделей и получения одной лучшей модели путем объединения соответствующего количества необходимых моделей, б) объединения достаточного количество моделей, наиболее близких к оцениванию всех прогнозов, чтобы их количество было меньше, чем в других объединениях. Также существует алгоритм, определяющий исследование, которое необходимо провести после наступления каждого события – расчет вероятности обоснования для отдельных моделей, а также парных и тройных моделей. Обсуждается задача использования этих алгоритмов в конкретной области – прогнозирование землетрясений и показаны результаты работы авторского алгоритма.

show abstract

Hydrodynamic and seismic response to teleseismic waves of strong remote earthquakes in Caucasus

Cited by 6 publications

References 28 publications

Earthquake forecast by imbalance machine learning using geophysical predictors

Earthquake forecast by imbalance machine learning using geophysical predictors

Earthquake Forecast as a Machine Learning Problem for Imbalanced Datasets: Example of Georgia, Caucasus

Анализ Моделей Прогнозирования Землетрясений Для Получения Наилучшей Модели

Contact Info

Product

Resources

About