1: \begin{abstract}

2: Motivated by Hubert's segmentation procedure \cite{Hub1997,Hub2000}, we

3: discuss the application of hidden Markov models (HMM) to the segmentation of

4: hydrological and enviromental time series. We use a HMM algorithm which

5: segments time series of several hundred terms in a few seconds and is

6: computationally feasible for even longer time series. The segmentation

7: algorithm computes the Maximum Likelihood segmentation by use of an

8: expectation / maximization iteration. We rigorously prove algorithm

9: convergence and use numerical experiments, involving temperature and river

10: discharge time series, to show that the algorithm usually converges to the

11: globally optimal segmentation. The relation of the proposed algorithm to

12: Hubert's segmentation procedure is also discussed.

13: 

14: \end{abstract}

15: