Cold-air damming (CAD) is a prevalent Mid-Atlantic United States weather phenomenon that occurs when cold, dense air is dammed alongside the eastern slopes of the Appalachian Mountains. Lower-than-normal maximum temperatures, increased and prolonged cloud cover, and precipitation that produces hazardous impacts are common features of this weather event, which are well known for presenting difficulties to both human forecasters and weather prediction models. This study explores CAD events between 2007 and 2016 archived in a Blacksburg National Weather Service ‘bust’ database – instances when forecasters erred by at least 8°F (4.4°C) on either maximum or minimum daily air temperature. The database includes the temperature error within Model Output Statistics (MOS) guidance in association with these forecast ‘busts.’ During the 10-year study period, MOS guidance produced warm-biased maximum temperatures and cold-biased minimum temperatures for most of the problematic CAD events, suggesting MOS guidance tended to underestimate the strength of CAD in these cases, seeming to struggle with weaker CAD events. During CAD erosion, MOS tended to prematurely erode CAD scenarios at night and predicted them to persist for too long during the day. Hourly surface meteorological and synoptic atmosphere composites during these ‘busted’ CAD events failed to reveal obvious differences from what is expected for central Appalachian CAD. However, a comparison to well-forecast classic cold-season CAD events suggest that busted cases of this same type of CAD may be drier than is typical. As the atmospheric patterns associated with busted CAD events are typical of the phenomenon, but a bit weaker or more marginal, forecast errors appear to stem from subtle model errors rather than forecaster error. It is possible that the models may inadequately characterize low-level moisture, but further research is needed to isolate the source of model forecast error. Nonetheless, the results of this research serve as guidance for operational forecasters as they consider model guidance during weak CAD events. / Master of Science / Cold-air damming (CAD) is a common weather pattern that affects the Blue Ridge Mountain region of the eastern United States, in which cold air at the atmosphere’s surface is directed from the Northeast and is dammed against the eastern Appalachian Mountains. This weather event causes lower-than-normal temperatures over the region and is often characterized by prolonged cloudy skies and precipitation. CAD is very difficult for forecasters to accurately predict, as they rely on weather forecast models that often simulate these situations poorly. CAD also strains emergency managers who rely on accurate forecasts to support public safety during CAD. This study explores CAD events between 2007 and 2016 archived in a Blacksburg National Weather Service ‘bust’ database – instances when forecasters erred by at least 8°F (4.4°C) on either maximum or minimum daily air temperature. The database includes the temperature error within Model Output Statistics (MOS) guidance in association with these forecast ‘busts.’ During the 10-year study period, MOS guidance forecasted maximum temperatures too high and minimum temperatures too low for most of the problematic CAD events, suggesting MOS guidance tended to underestimate the strength of CAD in these cases, seeming to struggle with weaker CAD events. During instances where CAD dissolved from the Appalachians, MOS tended to prematurely erode CAD scenarios at night and predicted them to persist for too long during the day. Hourly surface meteorological and synoptic atmosphere composites during these ‘busted’ CAD events failed to reveal obvious differences from what is expected for central Appalachian CAD. However, a comparison to well-forecast classic cold-season CAD events suggest that busted cases of this same type of CAD may be drier than is typical. As the atmospheric patterns associated with busted CAD events are typical of the phenomenon, but a bit weaker or more marginal, forecast errors appear to stem from subtle model errors rather than forecaster error. It is possible that the models may inadequately characterize low-level moisture, but further research is needed to isolate the source of model forecast error. Nonetheless, the results of this research serve as guidance for operational forecasters as they consider model guidance during weak CAD events.