A winter storm impacted much of the Carolina’s and Virginia with periods of locally heavy wet snow on Saturday 16 February, 2013. The primary area of snow across central NC was drive by strong low and mid-level frontogenesis, a strengthening 150kt upper level jet all within a region of strong mid-level height falls associated with deepening negatively tilted trough. While temperatures were cooling above the boundary layer in response to the approaching trough, there was limited cold horizontal thermal advection at the surface with cold high pressure well removed from the region. Forecast soundings suggested that the precipitation would likely fall as snow but melt, at least initially, with surface temperatures in the mid 30s to lower 40s. With this in mind, forecasters watched the thermal profiles and precipitation in the forecast soundings during the event very closely.
When the event drew close enough for the RAP to begin capturing it in its limited temporal window, some bizarre surface and boundary layer processes were observed in the RAP forecasts for KRDU. A similar tendency was noted at some other locations as well. In the Bufkit timeline display of the 2/16 15Z RAP BUFR forecast for KRDU shown to the right, the hourly 2-meter temperature forecast is shown in red, initializing at 36 degrees F at 15z, increasing to 37F at 16Z, 41F at 17Z, 43F at 18Z and remaining near that temperature for several hours afterward. During this time the “skin temperature” shown in light blue climbs into the lower 50s by 17Z.
This is all happening while hourly precipitation rates range between 0.02 and 0.07 through 19Z with a four hour total of 0.16 of liquid equivalent. Even more surprising, the thermal and moisture profiles suggest saturation well into the ice nucleation region with at least some upward vertical motion suggesting that the precipitation should be snow aloft. The snow will fall and then reach the top of a melting layer beginning at around 1,500 ft AGL. The snow would then melt in the boundary layer or at the surface and the melting would promote cooling near or at the surface.
The filmstrip to the right and below shows the lower portion of a skew-t display of the 2/16 15Z RAP BUFR forecast for KRDU beginning with the initialization hour (15Z) and continuing through F5 (20Z). Similar processes as described previously can be noted with this visualization. In addition the lower level flow remains northerly throughout the period and no obvious source of warm air was available to the north (note that other RAP BUFR forecasts for locations north of Raleigh such as Richmond VA showed a similar pattern of near surface warming).
The very warm surface temperatures result in an unrealistic surface based super adiabatic layer. The observed temperature at KRDU at 15Z was 34 degrees with a mix of rain and snow falling. The precipitation would change to snow within the hour and moderate snow was reported at KRDU for the next four hours through 19Z.
The image below is a four panel display of forecast soundings valid at 18Z on 16 February from the 13Z RAP in the upper left (12Z was not archived), the 12Z NAM in the upper right, the 12Z GFS in the lower left, and an observed 1833Z AMDAR aircraft sounding at KRDU. While the GFS looks similar to the RAP, the GFS 2-meter surface temperatures were cooling during the late morning and midday hours (surface temperatures initialized at 39F at 12Z fell to 35F at 21Z). During this same period the RAP 2-meter temperatures were significantly warming.
A snap shot of the hourly surface observations from KRDU from 09Z to 21Z is shown in the image below.