Evaluating the performance of individual-based animal movement models in novel environments

摘要: Abstract Simulating animal movement in spatially explicit individual-based models (IBMs) is both challenging and critically important to accurately estimating population dynamics. A number of different approaches have been developed that make assumptions about how individuals move their environment use mathematics translate cues into a behavioral response. Properly calibrated should produce realistic conditions encountered during calibration novel conditions; however, most studies date not tested conditions. We compared four distinct or sub-models (restricted-area search, kinesis, event-based, run tumble) using an IBM loosely based on small pelagic fish (e.g. Engraulidae) simulated growth, mortality, cohort 2-dimensional grid. trained the with genetic algorithm one set environmental then them other three environments. The generally performed well environments, except restricted-area search event-based needed be environments gradients similar test environment. Also, tumble produced near-random distributions all training steepest habitat quality gradient, it random In selecting sub-model, researchers consider potential sub-models, observed patterns species interest, shape steepness underlying gradient.