Keith Roberts led a discussion titled, “Coming to a sky near you: A chance of rain”
The weather discussion led of December 5th began with a visualization of the 10-m wind field around the Northern hemisphere to demonstrate a tool that can visualize and aggregate real-time observations. At the time of the discussion, it was shown that there was anti-cyclonic flow over the northeast region with near seasonable temperatures and broken skies. As such, there was weak subsidence at 700 mb as indicated by 700 mb q-vector divergence. The past 2 weeks had been somewhat active with a southern stream cyclongenesis event around Thanksgiving followed by the passage of a Canadian maritime front on December 1-2.

Focus was turned to the present conditions. The flow at 500 mb over the eastern United States was zonal and progressive. Between Saturday-Sunday, a cold front was forecasted to clear the Eastern seaboard and set-up a baroclinic zone for potential cyclogenesis off the southeast U.S. The location and timing of the cyclogenesis event was considerably uncertain in model guidance. Deterministic model guidance was first analyzed. The differences between model guidance bifurcated into two camps: UKMET/ECMWF/CMC (more progressive and more intense cyclogenesis) and WRF/GFS (slower and weaker cyclogenesis) in the timing and phasing of the northern stream shortwave that was associated with the passage of the Arctic front on December 6-7. A spaghetti tool put up by the WPC demonstrated large differences at the 500 mb in the timing and amplitude of the northern stream shortwave between the two aforementioned model camps at rather short lead times (< 36 h). However, the majority of modeling systems eventually did develop a cyclone off the southeastern US coast by mid-week (Dec 9-10), but the movement of this cyclone and its intensity (1000-1020 hPa MSLP) had tremendous variance. To illustrate the variance in cyclone tracks, the cyclone tracks for various ensemble guidance systems (SREF,GEFS, etc.) was shown with cyclones tracking anywhere between Alabama and Boston at similar times. The large uncertainty was hypothesized to be attributed to diabatic effects associated with the western bound of the Gulf stream wall as upper level QG dynamics did not appear to be largely present. To synthesize the large spread and uncertainty in model guidance, EOFs were analyzed for some of the GEFS and ECMWF guidance. The first EOF in both ensembles explained > 40 % of the ensemble variance in 500 mb. Day 4 (December 9th forecast) EOF1’s sensitivity could be traced back to a disturbance north of the Arctic circle in the Aleutian sea. At this lead-time (+4 days), the sensitivity analysis tool was best suited for identifying where the storm’s uncertainty was originating from rather than serving to be physically interpretable. This type of downstream sensitivity to disturbances north of the Arctic circle has been routinely observed in the northeast region since this tool has become operational. It motivates a more complete sampling of the Arctic weather conditions for a more accurate model initialization and theoretical work on the nature of uncertainty and its propagation in modeling systems. The largest sensible weather impacts were high wind on the backside of the Arctic front Sunday-Monday and heavy rain near the coast with a wet mix inland. The potential for coastal frontogensis mid-week as the cyclone moved up the coast was suggested by Dr. Brian Colle due to deformation between the cyclonic on-shore flow (maritime air) and the Arctic air inland associated with the anti-cyclonic over the north east. Due to the large uncertainty in the mid-week cyclogenesis event, it was difficult to say what the sensible weather implications would be, however, some guidance suggested a persistent on-shore flow in the NY/NJ Bight, which could lead to widespread coastal flooding.