Fig. 5

Log relative selection strength (log-RSS) from the A RSF on the full dataset, RSF on the thinned dataset (every 10th location), and SSF without movement-habitat interactions and B SSF with movement-habitat interactions, where prey diversity affected step length; log-RSS was calculated for speeds at the 25th (slow), 50th (moderate), and 75th (fast) percentiles of step length, which we infer as “speed” since the locations are recorded at a fixed time interval (see [26, 106]). C Estimated state probabilities from the HMM as a function of prey diversity (on the transition probability). Shaded areas represent the standard error. Note the bowtie shape of the standard error in A–B is due to log-RSS calculating selection strength relative to a starting step, in this case, where prey diversity is set to its mean