Fig. 3

Model validation comparisons between Bayesian geospatial model and generalized linear model (GLM). Predictions based on the Bayesian geospatial model generally outperform the GLM. Two out-of-sample metrics were calculated; the log-likelihood for each survey (a) and MAE (b). Grey vertical polygons depict the range of outcomes based on the 10 cross-validation folds. Higher values for the log-likelihood and lower values for the MAE indicate better out-of-sample predictive skill