The previously introduced algorithm of optimization recursive filtering allows separation of the signal and the harmonics in non-explosive seismic data. This opens up the possibility of separately constructing correlograms or stacks based on the main sweep signal and its harmonics of different orders. It turned out that summation of results in overlapping frequency ranges not only does not make sense, but can also lead to the appearance of specific artifacts. Correlograms and sections are constructed in non-overlapping frequency ranges with their subsequent stitching using the analytical spectrum extrapolation technique.

In this study the one-dimensional numerical modeling for sedimentary cover deposits of the Mansi syneclise in 17 wells was performed. Taking into account the calibration of the thermal history of the region, the density of the effective heat flow at the upper boundary of the Pre-Jurassic basement is estimated. It is shown that the highest values of the heat flow are confined to the central and southwestern parts of the syneclise, while the minimum values are confined to the southeastern part. The range of change in the average values ranges from 40 to 68 mW/m2.

This study investigates the impact of amplitude anomalies caused by heterogeneities in the near-surface layer on the accuracy of acoustic inversion results. The analysis is based on both synthetic data, generated via finite-difference modeling, and real seismic data. Inversion results demonstrate improved accuracy and stability after applying surfaceconsistent amplitude correction procedures. The study also investigates the reliability of models derived from seismic inversion under the influence of near-surface heterogeneity. The findings highlight the importance of evaluating the reliability of inversion-derived models when near-surface effects are present.

This article proposes and compares formalized criteria for the quality of automated well correlation. Three indicators are considered: the error in predicting well log values based on cross-validation; the ratio of the average intra-group variance to the total data variance; and the standard deviation of the depths of a set of stratigraphic boundaries from a reference model. A very close statistical relationship is demonstrated for the first two criteria, but due to the comparative simplicity of calculation, the second is preferable. The criterion associated with assessing the deviation from the reference model (created by an expert) can be used in machine learning for practical tasks, but is of little use because it requires the preliminary construction of a reference model.

The work is devoted to the study of the heterogeneity of the boundary of the Earth's crust and mantle in the central region of the Kamchatka Peninsula. For the research area, according to a dense network of 33 temporary and 14 permanent stations for 2019–2020, the values of the thickness of the Earth's crust and the average VP/VS ratio were calculated. To determine the depth of the crust-mantle boundary (Moho), the receiver function method was used, based on the analysis of the difference in the arrival time of direct and exchange phases of seismic waves from teleseismic events. The inversion of the receiver functions was carried out using the grid search method in the coordinates H (crust thickness) and k (VP/VS ratio), assuming that the crust is a homogeneous isotropic layer on a half-space. The results showed that the Moho depth ranges from 30–35 km under the stations in the Kamchatka central region, with a deepening to 40 km in the area of the Avacha Volcano. For the stations of the temporary network that were set up by the IPGG SB RAS staff, and most of the stations of the KB GS RAS permanent network, the result was obtained for the first time and in the intersection areas is consistent with the conclusions contained in previously published works by other authors.