Abstract: Modeling metapopulation dynamics is potentially a useful and powerful tool for ecologists and conservation biologists. However, the key processes driving metapopulation dynamics are colonization and extinction events, which are notoriously difficult to study. A recent global assessment of Adelie Penguin (Pygoscelis adeliae) breeding populations reported multiple potential colonization and extinction events from occupancy observations using recent high-resolution satellite imagery and examination of historical occupancy literature, but emphasized that the events should be confirmed with direct observations. We evaluated these conclusions by developing a comprehensive database of direct observations of Adelie Penguin breeding site occupancy across East Antarctica. The database allowed 16 of 19 proposed colonization and extinction events in this region to be evaluated using direct observations that were concurrent in space and time (same breeding season) with observations from satellite imagery. We concluded that none of the 16 proposed colonization and extinction events had occurred. One true extinction event may have been correctly identified from satellite imagery, but a conclusive evaluation was not possible because the direct and satellite observations were not concurrent in time. The evaluation identified several sources of error in observations from satellite imagery, including errors of omission through failure to detect small colonies and errors of commission through misidentification of other biological and physical features as Adelie Penguin guano. The occupancy database corrected issues of poor precision and accuracy in locations of historical occupancy data. Our results improve understanding of key parameters for Adelie Penguin metapopulation dynamics and facilitate improved interpretation of satellite imagery in the future.