The spring riverine thermal bar phenomenon is investigated numerically on an example of Lake Baikal, and the spread of pollutants coming from the Selenga River is forecast using the 2.5 D non-hydrostatic model in the Boussinesq approximation. This hydrodynamic model takes into account the diurnal variability of the heat fluxes on the lake surface and the effects of wind and the Earth's rotation. The results of numerical modeling show that the variability of the total heat flux over 24 h plays a significant role in the variation of the thermal bar movement rate that contributes to the rapid mixing of impurities entering with river water.