Low Predictability and Associated Forecast Challenges with the January 3-5, 2018 Nor'easter (346)

Tracey Dorian, IMSG at NOAA/NCEP/EMC, St. Louis, MO

Logan Dawson, IMSG at NOAA/NCEP/EMC

Geoffrey Manikin, NOAA/NCEP/EMC



Low Predictability and Associated Forecast Challenges with the January 3-5, 2018 Nor'easterTracey A. Dorian, IMSG and NOAA/NCEP/EMC, College Park, M, L. C. Dawson, G.S. ManikinThe January 3-5, 2018 nor'easter rapidly intensified into a 950mb extra-tropical system producing blizzard conditions along much of the US east coast from Virginia to Maine with some significant snow accumulations. Rare winter precipitation in the form of sleet and freezing rain was also seen in southern areas like in Georgia and Florida. The Model Evaluation Group (MEG) at NCEPs Environmental Modeling Center (EMC), which evaluates model performance on a weekly basis and presents the results to the forecasting and modeling communities, examined this particular case in detail. The MEG found that the NCEP high-resolution models handled the timing of the intensification well, but overestimated the drop in mean sea-level pressure. The coarser models deepened the storm too late when compared to actual observations. This case was particularly challenging for the models because of vigorous convection off of the east coast and observed multiple low pressure centers. There was low predictability in the medium-range forecasts because of the difficulty in the models to predict the 500mb height patterns. In the short-range, the models largely underestimated the snow amounts, especially the global models because of their persistent eastern track bias - even just 24 hours before event time. The NAM and the NAM-nest models handled the track better than the global models, forecasting a track much closer to the coast. Because of the better track forecasts, the NAM and NAM-nest better captured the precipitation than the global models on the northwest side of the storm. As for ensemble performance, the 21 GEFS members similar to the deterministic GFS exhibited an east bias, the recently implemented HREF displayed skill in its precipitation-type probability forecasts, while the SREF spread was overly-dispersive.