Appalachian cold-air damming (CAD) can be accompanied by freezing rain, sleet, reduced ceilings and visibility, and below-average temperatures. Operational forecasters in the Appalachian damming region have listed CAD as a major forecasting challenge. In recent years, numerical weather prediction (NWP) models have proven fairly effective in the representation of CAD onset, but forecasters note a systematic tendency in the models for premature erosion of the CAD cold dome. This research is designed to (i) elucidate the physical processes responsible for CAD erosion in various synoptic settings, and (ii) examine the ability of NWP models to represent these processes.

Several of the physical mechanisms of CAD erosion can be summarized by consideration of the bulk Richardson number applied to the inversion layer atop the CAD cold dome. Processes that weaken the inversion (reduce Ri) render the cold dome more susceptible to mixing; increases in cross-inversion shear can also promote mixing if Ri is sufficiently reduced. Examination of the demise stage of 90 CAD events indicates that cold advection aloft frequently serves to weaken the inversion, especially in scenarios characterized by cold-frontal passages either at the surface or aloft, or the development of coastal cyclones. Warm advection near the surface, which may accompany warm/coastal front passage, can also serve to weaken the inversion from below. However, near-surface warming associated with solar heating is a more common mechanism for bottom-up CAD erosion. Finally, divergence within the cold dome serves to reduce the depth of the cold air, and in certain synoptic settings can promote the inland progression of the coastal front.

Several aspects of model representation of CAD erosion processes are investigated using a case study from October 2002, including (i) cloud-radiation interaction, (ii) PBL scheme representation, and (iii) representation of land-surface processes. These results indicate that adequate model representation of CAD erosion in this case is dependent upon accurate representation of the synoptic-scale thermal advection pattern. Premature CAD erosion in numerical forecasts of this event appear linked to synoptic-scale thermal advection patterns, overestimation of surface solar radiation, and PBL parameterization errors.