The next issue of The Russian Journal of Seismology (Vol. 6, No. 3) has been published. The issue contains articles written by scientists from different regions and research organizations of Russia.

Traditionally, at the beginning of each half-year, our journal publishes a regular article “Global earthquakes in the first half of 2024 according to the GS RAS” (authors: Yu.A. Vinogradov, M.I. Ryzhikova, N.V. Petrova, S.G. Poygina, M.V. Kolomiets), which presents information on the seismicity of the Earth in the first half of 2024 at a magnitude level of mb (MS) ≥ 6.0 (a total of 60 earthquakes), as well as information on 78 earthquakes felt on the territory of the Russian Federation according to the Alert Service of the Geophysical Survey of the Russian Academy of Sciences. It is noted that the seismic energy released on the globe in the first half of 2024 increased compared to that in the second half of 2023, but remained below the average semi-annual value for the period 2010–2023, as in the two previous years.

Representatives of the Seismological Branch of the GS RAS from Novosibirsk I.V. Kokovkin and V.S. Seleznev applied their developments to assessing the safety of train traffic in the Novosibirsk metro. The article "Remote control of the technical condition of metro trains using seismological tools " proves that a seismic recorder installed in a house next to a metro station can be used to monitor changes in the amplitude-frequency spectrum of a vibroseismic record. The need to develop software equipped with a database containing information for past months and years is stated, which will allow comparing current indicators with past ones and promptly responding to significant deviations, thereby increasing passenger safety.

Our colleagues from the Baikal region (Ulan-Ude and Irkutsk) P.A. Predein, M.A. Khritova presented the article "Seismic moment of local earthquakes in the central part of the Baikal rift by the coda envelope inversion". This publication considers the determination of the properties of the radiation source, the value of the scalar seismic moment of earthquakes, using the coda envelope inversion method. It is shown that a simplified model of shear wave scattering can be used for energy classification and determination of focal parameters of weak and moderate earthquakes recorded with a good signal/noise ratio only at local distances.

A team of authors from several research and educational organizations in Moscow (N.I. Frolova, N.S. Malaeva, S.P. Sushchev) demonstrated in the article “Simulation of January 22, 2024 Mw=7.0 Uqturpan earthquake consequences with the “Extremum” system application” the results of taking into account regional features of seismic intensity attenuation in the Tien Shan territory when modeling the consequences of the Uchturfan earthquake on January 22, 2024. For the first time, an analysis of the applicability of seismic intensity attenuation equations obtained by researchers in different years and in different countries was performed for the territory under consideration for the rapid assessment of possible consequences of strong events. The results of modeling the consequences of the Uchturfan earthquake are presented, and an assessment is given of the convergence of the calculated and observed intensities for various intensity attenuation equations obtained earlier for China, Kyrgyzstan and adjacent territories in the area of ​​the Gissar-Kokshaal fault of the Tien Shan.

For the first time, the Russian Seismological Journal raises the issue of using machine learning methods to identify the nature of seismic events. The article "Convolutional neural networks and seismogram fingerprints as a tool for recognizing the nature of seismic events" (K. Yu. Silkin) demonstrates two facts. First, despite their brevity, fingerprints are informative enough representations of signal records to carry information about the nature of the registered seismic event. Second, it is shown that it is practically possible to design and train an artificial neural network capable of classifying events by origin based on their fingerprints with high accuracy. The convolutional neural network prepared for working with fingerprints has one of the simplest convolutional neural networks and a very modest number of adjustable parameters. Thanks to it, the classification accuracy of 95% of seismic events was easily