Observation of marine animals in their environment – whale-watching – has grown greatly in recent years, bringing risk to the animals. Of particular concern are harmful impacts on marine mammals, some of which are endangered. As a result, regulations have been developed for their protection, but these conservation measures require enforcement across a broad geographic region, which is difficult due to limited monitoring resources. A ship-borne information transmission system called AIS – Automatic Identification System – can provide information-rich marine vessel movement data that can be used to passively monitor vessels engaged in viewing wildlife, aiding regulatory bodies with compliance enforcement. Few studies explore the use of AIS data to determine when vessels are engaged in wildlife-viewing, and as such little guidance exists on how to implement classification models appropriately. The objective of this thesis is to use AIS data to evaluate the accuracy and utility of existing classification models to detect vessels engaged in observing wildlife, and determine whether information about species being observed can be extracted. Using a control set of observed cetacean encounter data, three classification models were statistically assessed. From this, a hidden Markov model was chosen for detailed analysis in the vicinity surrounding Vancouver Island, B.C., Canada. The resulting analysis concluded that a hidden Markov unsupervised classification approach was feasible for detecting vessel behaviours and differentiating species type. These findings suggest AIS can aid managers and the commercial whale-watching industry in making informed decisions regarding conservation regulations and their compliance. / Graduate / 2022-08-12