Helioseismology has revealed many details of solar differential rotation and its time variation, known as torsional oscillations. So far there is no generally accepted theoretical explanation for torsional oscillations, even though a close relation to the solar activity cycle is evident. On the t... Show moreHelioseismology has revealed many details of solar differential rotation and its time variation, known as torsional oscillations. So far there is no generally accepted theoretical explanation for torsional oscillations, even though a close relation to the solar activity cycle is evident. On the theoretical side, nonkinematic dynamo models (including the Lorentz force feedback on differential rotation) have been used to explain torsional oscillations. In this paper we use a slightly different approach by forcing torsional oscillations in a mean field differential rotation model. Our aim is not a fully self-consistent model, but rather to point out a few general properties of torsional oscillations, and their possible origins, that are independent from a particular dynamo model. We find that the poleward-propagating high-latitude branch of the torsional oscillations can be explained as a response of the coupled differential rotation/meridional flow system to periodic forcing in midlatitudes of either mechanical (Lorentz force) or thermal nature. The speed of the poleward propagation sets constraints on the value of the turbulent viscosity in the solar convection zone to be less than 3 × 10⁸ m² s⁻¹. We also show that the equatorward-propagating low-latitude branch is most likely not a consequence of mechanical forcing (Lorentz force) alone, but rather of thermal origin due to the Taylor-Proudman theorem. Show less