The main challenge in the prediction of tectonic earthquakes and their control is still insufficient awareness of seismotectonic processes in the lithosphere and upper mantle during the preparation of strong earthquakes. This is associated in many respects with not quite appropriate equipment for researchers. Among relevant problems is also the lack of adequate models of preparation of earthquake sources at different stages, and this retards the development of earthquake predication methods. The paper discusses long-term research on deformation and destruction of the Baikal ice cover in the context of physical mesomechanics. With certain combinations of meteorological factors (wind, temperature, precipitation, undercurrents, etc.) responsible for deformation, major cracks of many kilometers arise in the Baikal ice cover. Their spontaneous growth often involves seismic phenomena as ice quakes whose energy reaches Emax 104-107 J. The nucleation of major cracks is similar to that of rock bursts of moderate strength or weak earthquakes. It is found that ice quakes and earthquakes are both preceded by foreshocks, seismic calm for tens of minutes, aftershocks and other events against the background of accelerated creep in fractures and increased seismoacoustic activity. Research data make it possible to put forward two genetically interrelated criteria among basic factors for ice shock prediction: variations in deformation modes at convergent boundaries of ice plates and a specific intensification mode - generation of strong foreshock pulses in a segment in which ice plates are prepared for dynamic motion. We substantiate the conclusion that simpler and clearer scenarios of preparation of strong seismic events in the Baikal ice cover allow successful physical simulation of preparation of tectonic earthquakes and rock bursts and advances in their prediction. We also consider and substantiate the feasibility of techniques for more efficient seismic risk reduction
