Understanding narwhal diving behaviour using Hidden Markov Models with dependent state distributions and long range dependence.

摘要: Diving behaviour of narwhals is still largely unknown. We use Hidden Markov models (HMMs) to describe the diving a narwhal and fit three-dimensional response vector maximum dive depth, duration dives post-dive surface time 8,609 measured in East Greenland over 83 days, an extraordinarily long rich data set. Narwhal patterns have not been analysed like this before, but studies other whale species, variables assumed independent. extend existing allow for dependence between state distributions, show that has impact on conclusions drawn about behaviour. try several HMMs with 2, 3 or 4 states, independent dependent log-normal gamma respectively, different covariates characterize patterns. In particular, diurnal inferred, by using periodic B-splines boundary knots 0 24 hours.