1/Q² terms and Landau singularity

Standard power-behaved contributions in QCD arising from non-perturbative effects at low scale can be described, as shown by Dokshitzer, Marchesini and Webber, with the notion of an infrared regular effective coupling. In their approach, a non-perturbative contribution to the coupling, essentially restricted to low scales, parametrizes the non-perturbative power corrections. I argue that their framework naturally allows for another type of power contributions, arising from short distances (hence unrelated to renormalons and the operator product expansion) which appear in the process of removing the Landau singularity present in perturbation theory. A natural definition of an infrared finite perturbative coupling is suggested within the dispersive method. Implications for the tau hadronic width, where $O(1/Q^2)$ contributions can be generated, are pointed out.