Emelie Siderud, Kjell Eriksson, Susanne Höfner, Arief Ahmad

Mass loss from asymptotic giant branch (AGB) stars is the result of a complex interplay between pulsation, atmospheric dynamics, dust formation, and radiative acceleration. Pulsation periods are a key input in dynamical atmosphere and wind models, and different prescriptions for assigning periods based on stellar parameters may lead to systematic differences in the predicted wind properties. To better constrain this critical parameter, we investigated how the choice of pulsation period affects the wind properties of dynamical atmosphere and wind models of M-type AGB stars by comparing models based on an empirical period-luminosity (P-L) relation with corresponding ones that adopt a period-mean density relation derived from 3D pulsation models. We analysed two grids of DARWIN models that cover a range of current stellar masses, luminosities, and effective temperatures. For each grid, pulsation periods were assigned using either the P-L relation or the period-mean density relation, allowing for a direct comparison of the resulting dynamical structures and wind properties for pairs of models differing by period only. Independent of the adopted period prescription, the time-averaged wind properties correlate strongly with $L_\star/M_\star$. The pulsation period affects the atmospheric dynamics through changes in the relative timing of shock propagation and dust formation, which affect both wind formation and the resulting wind properties. Shorter periods favour the onset of a wind, and models differing only in pulsation period can exhibit significantly different wind properties. The period-mean density relation provides a physically motivated alternative to the empirical P-L relation by accounting for stellar parameters beyond luminosity, and enables a more direct comparison between DARWIN models and observed Mira variables.