This work concerned three-dimensional gravity modeling of the Gorodishche gabbro-anorthosite massif located within the Korsun-Novomirgorod pluton of the Ukrainian Shield. A three-dimensional model of the upper crust was created using maps of the anomalous gravitational field at a scale of 1:200 000, taking into account data from seismic methods of various modifications. The differences in the structure of the intrusive complex and its gneisses were reflected in seismic wave fields, allowing us to determine the boundaries of the entire intrusive massif: rapakivi granites and basic rocks. Three-dimensional gravity was modeled to separate these complexes of different densities. This made it possible to identify gabbro-anorthosite bodies with a maximum thickness of 5 km in the upper part of the section, and to study the contacts of the intrusive complex and gneisses. In the process of solving inverse problems, various criteria for local optimization of gravitational field sources were implemented. In the iterative process, three different functionals were calculated. The mean square of the difference between the observed and theoretical fields is F1, the average sum of the moduli of the field difference is the functional F2, and the preliminary logarithm of the parameters is the functional F3. The joint use of functionals makes it possible to reduce various types of noise in the observed potential field data. In the process of research, it was found that the use of functionals of various types in the selection method algorithms is quite advisable. This makes it possible to create new algorithms that improve feasible solutions to the inverse problem, which is necessary for the implementation of additive technologies for the quantitative interpretation of gravity survey data. The constructed model, which takes into account all available a priori information about the density and geometric parameters of anome-forming objects, can be used to obtain additional reliable geological information about gabbro-anarthosite massifs.
Read full abstract