ИСТИНА

Recent years have been marked by the appearance of publications dedicated to numerical models of convective interaction between subduction zones and adjacent domains of the upper mantle. The essence of such interaction consists in the formation of a temporally changeable hydrodynamic system by convective cells, which join the upper mantle on both sides of the subducted plate. The external sides of these cells remote from the subduction zone are always marked by the formation of ascending mantle flows, which compensate subductioninduced subsidence. These flows may be considered as upper mantle plumes. In the hydrodynamic model, subduction serves as a triggering mechanism for upper mantle plumes and such a causeandeffect relation is observable in many Pacific and Mediterranean regions. The mutual effect of Pacific subduction and the adjacent convective cell is considered as a mechanism responsible for evolution of the Arctic lithosphere in the Mesozoic–Cenozoic. The horizontal migration of island arcs toward the ocean accompanied by the formation of backarc basins is the most important feature of the subduction process. For example, the Izu–Bonin subduction zone migrated in the northeasterly direction for a distance of 1000 km in the period of 30 to 17 Ma ago. At the same time, it is known that some regions are characterized by reversed migration of the subduction zones toward the continent; moreover, such a horizontal shift of the subduction zone may reach 80–210 km and its velocity may be as high as 60 km/10 mln years. The deepseated physical mechanism of such a migration of island arcs together with subduction zones remains unclear. As follows from numerical experiments, the hydrodynamic model under consideration including the subduction zone and adjacent convective cells is characterized by a nonstationary regime with cyclic development phases. It is logical to correlate cyclic geological processes with this phenomenon. The issue is that the Phanerozoic history of the Earth includes approximately 20 orogenic phases. V.E. Khain suggested to call them the Stille cycles in contrast to the longer Bertrand cycles (150–200 Myr). The orogenic cyclicity is evident from reorganizations of the direction and velocity of motion of lithospheric plates, jumplike displacement of spreading and subduction zones, irregular ophiolite and granite formation, volcanism, and metamorphism. Thus, in this communication, we suggest to consider the problems of geological cylicity in the context of nonstationary subductioninduced processes. The numerical models of the subduction process developed by many researchers concern isolated regions, i.e. when the influence of adjacent domains on subduction is excluded. The authors postulated the absence of forces as well as heat and mass flows at the boundaries of the region (specularly symmetrical boundary conditions), which means compete isolation of the region and its exclusion from the general geodynamic process. Such a simplified formulation of the problem is inadequate since conditions on boundary surfaces provide a strong influence on the hydrodynamic regime of the process. We faced this problem when modeling the geodynamics of the Arctic region revealing that simplified formulation of the regional problem cannot explain its evolution. In this investigation, we used a different approach to numerical modeling of regional processes, which is based on strict mathematical methods of solution of illconditioned problems and our previous data on global mantle convection. The regional modeling of nonstationary convection in the upper mantle with the improved formulation of the problem allows the abovementioned features in the behavior of subduction zones and cyclic regularities to be reconstructed.