A modern method for discrete underwater channel characterization is the passive tomography concept which takes advantage of the generated signals by the natural sources (opportunity sources). The main difficulty which appears in this field is due to the lack of any information about the received signals. In this paper, we present a method for underwater channel characterization, supposing that the received signal is a sum of shifted chirps, having the same chirp rate and start frequency. Hence, using the fractional Fourier transform, it is possible to express the received signal as a sum of sinusoids whose frequencies are directly related to the times of arrival. In order to distinguish the closed arrivals, we apply a high resolution spectral estimation method, providing also the time of arrivals.