Аннотация:The paper presents a methodology for building simulative physical-geological models of the near-surface based on the results of high-resolution seismic data interpretation. The main purpose of creating such simulative models is to quantify possible complications during offshore well drilling. In the future, the simulative models can be used to calculate synthetic seismic wavefields and train machine learning algorithms, including neural networks, aimed at identifying and assessing geological risks from seismic data.