Fig. 5

Results of three different segmentation methods using the simulated movement data. a The left panel shows the distribution of the observed step lengths as well as the applied cut-off value (threshold = 2 units). The proportions of the resulting behavioral states (short- and long-range movements) within and outside of the habitat are shown in the right panel. b Results from the behavioral change point analyses applied with the net-squared displacement signal. The observed time-series was segmented at significant change-points (vertical lines) to distinguish movements within the main ranges of the animal and two migratory periods. The color of the estimated parameter ρ^ indicates the level of temporal autocorrelation. c Change in switching probabilities between the two states (resting vs. active) dependent on the different hours of the day. Switching probabilities also differed with regard to whether the animal was in its habitat or not. Black lines indicate the switches from the resting state to the active state. Red lines are showing the switching probabilities from active to resting state