Fig 1.
Predicted and projected white abalone relative mean habitat suitability based on historical and contemporary distributions.
Upper panels represent suitability in the Southern California Bight (outlined in the solid bounding box of panel f) based on fishery-dependent (a) and fishery-independent (b) models during their respective time periods. Lower panels represent future predictions of relative mean habitat suitability throughout California waters (outline in the dotted bounding box of panel f) in 2050 under the RCP 2.6 scenario, based on the fishery-dependent (c) and fishery-independent models (d). Suitability interpreted from log-transformed relative mean abundance using random forest for the fishery-dependent model and relative probability of presence using MaxEnt for the fishery-independent model where unsuitable = zero and most suitable = 1 (or for fishery-independent model most suitable = 0.75). Relative mean habitat suitability is visualized for fishing blocks ≤ 500 m in depth. Note differences in suitability scales across predicted and projected panels.
Fig 2.
Response curves for broad-scale environmental variables.
Response curves for the fishery-dependent model (a-d) are developed from random forest and response curves for the fishery-independent models (e-h) are developed from MaxEnt. The black line represents the mean response of 100 model runs and the grey shaded region represents one SD of the mean.
Fig 3.
Predicted relative mean probability of white abalone presence (relative mean habitat suitability).
Predicted relative mean probability of white abalone presence in five areas located within the Southern California Bight (f), including San Clemente Island (a), Santa Catalina and Santa Barbara Islands (b), Tanner and Cortes Banks (c), San Diego (d), and Palos Verdes (e). Each panel represents the mean probability of abalone presence averaged over 100 model runs, ranging from unsuitable (0) to most suitable (0.7).