PARENT SESSION
Poster Session # 2: Statistical Ecology and Ecological Modeling.
Monday, August 4 Presentation from 5:00 PM to 6:30 PM. SITCC Exhibit Hall B.
Parameter estimation for the behavioral algorithms of a fish movement model.
Lemasson, Bertrand*,1, Haefner, James1, Bowen, Mark2, Frizzell, Kathy2, 1 Utah State University, Logan, Utah, U.S.A.2 Bureau of Reclamation, Denver, Colorado, U.S.A.
ABSTRACT- An individual-based movement model was statistically validated using extraction efficiency experiments conducted in the Tracy Fish Collection Facility (TFCF) in Tracy, CA. This work is part of the ongoing conservation efforts of the state of California and the Bureau of Reclamation in reducing the loss of endangered and economically important fish species in the lower San Joaquin river. The model incorporates individual fish behavior and swimming physiology with simulations of fluid flows in a physical model of the TFCF. Previous work used existing literature values for many behavioral parameters. Predicted extraction efficiency for 6 common species in the TFCF ranged between 0.66 and 0.83; sensitivity analysis quantified the effect of initial fish energy level and fish avoidance behavior on predicted efficiency. Model predictions were within the experimental error for most species, although statistical variability is high for measured efficiency. To more precisely test and parameterize the model, new validation and parameter estimation experiments were conducted using digital video taping of fish behavior in an experimental flume constructed by the Bureau of Reclamation, Denver, Colorado. The flume is a plexiglas chamber with an array of vertical louver slats and controls to vary bypass velocity ratio. We tested Rainbow Trout(mean size 98 mm) and Splittail(mean size 151 mm) using two approach velocities and louver configurations. Using images recorded every 0.1s and video tracking software, we estimated parameters of fish obstacle avoidance behavior from frequency distributions of turning angles and accelerations. Avoidance behavior is a change in the variance of these two parameters between time steps. To investigate the effect of different types of obstacles on fish behavior, we compared avoidance behavior near the louver array and near the opposite solid wall. We report these values and the results of extraction efficiency simulations based on new flume fluid dynamics and the refined parameter estimates.
Key words: fish behavior, Individual based model, parameter estimation