Analysis of a cold-air precipitation event: Observational diagnosis and numerical model sensitivity

The nature of a cold-air precipitation feature that developed prior to and upstream of the 24–25 January 2000 East Coast snowstorm is investigated from both an observational and modeling perspective. Previous research has shown that this area of precipitation initially generated a lower-tropospheric potential vorticity (PV) maximum important to downstream moisture transport into the Carolinas and Virginia later in the cyclone event. Despite a strongly stable surface airmass over the Gulf coast states, radar echoes and surface observations of thunder suggest that convection was occurring within this precipitation feature. Even though this precipitation formed over the relatively dense data network in the southern U.S., this feature was poorly forecasted in the 6–12 h range by operational numerical weather prediction (NWP) models. These same model forecasts later exhibited an easterly bias in the surface cyclone track and failed to produce heavy precipitation sufficiently far inland over the Carolinas and Virginia.

An examination of observations and model analyses indicate that this precipitation formed in a region of strong forcing for ascent associated with an upper-trough-jet system and a strong lower-tropospheric frontal zone. Above the stable surface airmass, conditional gravitational instability and conditional symmetric instability (CSI) were present, consistent with the observational evidence which suggests that elevated gravitational and/or slantwise convection were occurring. It is hypothesized that operational model forecasts of this precipitation feature were poor in part because the models were unable to properly resolve or parameterize the elevated and/or slantwise convection.

The Betts-Miller-Janjić (BMJ) convective parameterization (CP) scheme used in the operational Eta model at the time of this event the only examined the lowest 130 hPa above the model surface for instability, therefore the deep convection scheme would not activate if instability was located above this level. Additionally, slantwise convection is not parameterized in any operational model, and at the time of this event the Eta model was run with a horizontal grid-spacing of 32 km, which is unlikely to properly resolve this phenomenon on the grid scale. To test the above hypothesis, the ability of a mesoscale NWP model to simulate the formation of this precipitation will be tested with high-resolution model forecasts designed to examine the sensitivity to model initial conditions, CP scheme choice, and grid-spacing.