In this paper, the problem of identifying variations in the nature of atomic clock noise is addressed. Two methods are proposed. One method is based on a generalized likelihood ratio test (GLRT), and the other is based on the dynamic Allan variance (DAVAR), which is a representation of the instantaneous clock stability that is able to point out possible nonstationary behaviors. Both methods efficiently track variations in the experimental clock data and, thus, appear as suitable tools for the detection of atomic clock anomalies