This article presents a channel state information acquisition approach based on a Markov chain process that exploits information from the physical environmental conditions, including the tide phase and flow. The method is intended to predict channel characteristics, including the gain, delay, and Doppler spread, as well as the standard deviation of intrapath delays in time-varying conditions. Specifically, the correlation between different oceanic processes and the acoustic channel characteristics is confirmed to define a set of tide-dependent states corresponding to a particular channel condition. Channel soundings from a 34-day sea trial conducted in Grand Passage, Nova Scotia, are used to derive the channel characteristics statistics. For this purpose, channel soundings measurements are applied to a parametric model of the propagation channel. The probabilistic parametric model forms a data set by characterizing the time-varying channel impulse response and describing the channel tapped-delay structure statistically as a function of different tide phases. The proposed Markov chain is driven by the measured channel data set and predicts the future channel characteristics one tide cycle ahead. To validate the accuracy of the proposed method, the predicted channel characteristics are compared to the channel measurements obtained in a 566-m channel in Grand Passage, Nova Scotia.