[en] In this paper a number of phenomenological issues are identified for which the uncertainties are judged to be particularly important for sever accident analysis. The criterion for selection is based on considerations of contribution to uncertainty in risk and likelihood that research can reduce the uncertainty. Two containment loading issues and four source term issues are considered. The topics discussed include gas combustion (e.g., detonation and/or deflagration of high temperature hydrogen-steam-air mixtures), radioisotope release (e.g., source term from molten debris/concrete interactions), radioisotope transport (e.g, influence of radiation on iodine/air/water chemistry), and interactions of molten debris with containment atmospheres (e.g., aerosol production from dispersed debris droplets). In each of the areas the impact of the uncertainties on accident analysis is addressed and current research efforts (if any) to reduce the uncertainties are identified. 33 refs., 1 fig