The fast and accurate identification of unknown releases helps in an effective implementation of emergency preparedness (including plume exposure assessment) and public safety programs. The identification is based on networking of several receptors, limited a priori information, remotely measured concentrations, use of dispersion model, site description, meteorology, etc. This is considered as an ill-posed inverse problem. An overview of recently proposed/utilised inversion techniques is discussed with their various applications ranging from local scale to regional scale, limitations and inter-comparisons. The studies addressing identification of multiple-point releases are also included. In addition, the current issues, limitations and future scope associated with the inverse modelling are mentioned.