A numerical investigation of Appalachian cold-air damming with a WRF-based four dimensional data assimilation system

Cold air damming (CAD) is a weather phenomenon when cold air trapped by a major mountain range in a similar way as a dam blocks a river. In the United States CAD events commonly occur in the east bank of the Appalachian Mountains during winter and early spring when a parent high-pressure system settled over the Great Lakes or nearby range. The northeasterly flow from the anticyclone advects cold air to the east side of the Appalachians; it then rises up the eastern slope further cooling adiabatically (Bell and Bosart, 1988). Diabatic cooling from the melting and falling of associated precipitation may cause additional cooling (Xu, 1990). The Appalachian Cold-Air Damming often bring in severe weather like ice storm to the east region of the Mountains including Washington, D,.C. and Philadelphia. The climatology of the Appalachian CAD has been well documented (Bell and Bosart, 1988; Bailey et al., 2001 ). Xu and Gao (1995) used an analytic model to study the dynamical mechanisms involved in Appalachian Cold-Air Damming evolution.

Despite advanced understanding in the CAD event and continually improving mesoscale models, which now assimilate plenty of data with higher resolution than ever before, forecasting cold air damming has long been a great challenge. Numerical weather prediction models often depict synoptic signatures of damming reasonably well, but frequently fail to accurately detect small-scale features and processes which so heavily impact CAD development and evolution (Kramer, 1997). In this study, we revisit the numerical modeling aspects of the Appalachian CAD by using the NCAR WRF-based Real-Time Four Dimensional Data Assimilation (RTFDDA) system (Liu et al., 2008). With the RTFDDA system, and available observational data, a number of experiments have been conducted aiming to understand relative role and deficiency of different numerics, dynamics, physics and initial and boundary conditions in the model system. Our preliminary analysis indicates that 1. large-scale forcing can be important in positioning the parent high-pressure system that provides the cold air source; 2. errors in lower boundary conditions (e.g., coastal area sea surface temperature) is a contributor to near surface warm bias; 3. model physics and parameterizations in the stable boundary layer plays a critical role in air mass lofting and associated adiabatic and diabatic cooling; 4. modeling low-level blocking effect of the Mountains is also important to CAD intensity.