Several mode identification methods have been proposed in the literature to estimate oscillatory modes contained in power grid ringdown responses. Prony, eigenvalue realization,
matrix pencil and vector fitting are a few of them. In this presentation a systematic evaluation of their performance is conducted with special emphasis on the identification of closely spaced in frequency and poorly damped modes. Synthetic signals distorted by white Gaussian noise to replicate real-world measurement data have been used. The accuracy of the mode estimates is evaluated by using Monte Carlo simulations. The proposed synthetic signals can provide a reference for testing and validating the efficiency of mode identification algorithms.