Connecting models to movements: testing connectivity model predictions against empirical migration and dispersal data