Fig. 6

Performance of each of the models (on the y-axis) in predicting the number of prey encounter events (nPEE) at the scale of dives (A and B) and days (C and D). Each of the models is fitted with a generalized linear mixed-effect model (GLMM; in black) and with a boosted regression tree (BRT; in grey). The root-mean-square error (RMSE) and the correlation coefficient are computed between the observed values of nPEE for each seal and the predicted values by the model fitted without the focal individual (i.e., leave-one-individual-out cross-validation). Dive metrics are calculated from the low-resolution dive profiles, i.e., simplified to five segments using the broken stick method similar to data transmitted by CTD-SRDL loggers