The origin of decadal variations in the excitation of polar motion occurs to be one of the remaining open questions in Earth rotation studies. SinceWilliam Markowitz discovered the multi-decadal wobble that has been named after him, geodetic science has been in search of the possible underlying physical mechanisms for it. Although a combination of different processes is the most likely scenario and most studies suggest processes in the core to account for the main contribution, a complete picture of the whole phenomenon is still missing. Atmospheric processes, although of subordinate magnitude, also take part in decadal polar motion excitation. The present study investigates this decadal-scale atmospheric excitation over the whole twentieth century by using meteorological data from two different reanalysis systems. On one hand the thesis estimates the atmospheric contribution to decadal-scale wobbles by comparing geophysical excitation measures to geodetic observations of polar motion variations. On the other hand two reanalysis models are tested for their rational skill and consistency trough the angular momentum budget equation, i.e., the mathematical framework that is the foundation of a reliable estimation of the atmospheric contribution. In the end, a objective judgement on the usability of the reanalyses for Earth rotation studies is given, and the possible superiority of one of the probed datasets is pointed out. The atmospheric contribution is found to be small but not negligible. Beside that, good results in the angular momentum budget check justify the usage of both reanalyses models.