Fig. 6
Results from generalized additive models with a negative binomial error distribution assessing environmental variables affecting Alligator Gar (Atractosteus spatula) kurtosis values. Data are subsetted such that kurtosis values originate from dispersal data of filtered individuals from the M interval (i.e., 1–60 day relocation intervals). Each observation denotes the kurtosis value for a given tracking event for the Brazos (BR; circles) or Trinity (TR; triangles) rivers. In (a), the z-score transformed 30-day average daily discharge (Qz−30d) is plotted against kurtosis values while in (b) the 30-day average daily temperature (T°C−30d) is regressed against kurtosis values. Predictor variables were calculated using the previous 30 days from the last day in each tracking event. Solid lines exhibit a significant relationship (p < 0.05) while dashed lines resemble non-significance (p > 0.05). Gray shaded regions for each line display 95% confidence intervals