Ad-hoc radiative damping

If a TAE mode is strongly localized within a toroidicity gap and the peaking of the wavefield such that mode conversion becomes possible where $k^2_{\perp,\mathrm{TAE}} \approx k^2_{\perp,\mathrm{KAW}}$, the amount of power ``radiated away by the kinetic-Alfvén wave'' can be calculated perturbatively directly from the shear-Alfvén wavefield [32,33]. Assuming that all the power is finally absorbed in the vicinity of the conversion region (excluding reflections and fast particle drive on the kinetic Alfvén wave), the so-called radiative damping has been evaluated with the NOVA-K code. Choosing a KAE in JET where the gyrokinetic PENN code a priori predicted that this particular conversion / damping mechanism is dominant (Fig.1 in Ref.[31]), the radiative damping and the electron Landau damping obtained from a self-consistent gyrokinetic description of the global wavefield [35,31] are found to agree within approximatively 20% [31,35]. The problem with an ad-hoc evaluation of radiative damping is that it is not a priori possible to know if other conversion mechanism exist and dominate; choosing another KAE where the PENN code predicts that conversion takes place because of weak magnetic shear in the core, order of magnitude discrepancies appear between the damping from the two codes [35,31]. An ad-hoc radiative damping model is for example unable to reproduce the isotope scaling from Ref.[11], for which good agreement has been achieved between PENN and the measurements from JET.