Eastern United States Winter Storm of 12-14 February 2014 Dealing with Divergent Model and Ensemble Forecast Systems

1. 
   Introduction
   

The concept of the storm was relatively well-predicted by most of the major modeling systems. Additionally, the snow and ice in the southern United States was relatively well-predicted. However, the NCEP GFS and GEFS predicted the storm and the associated precipitation shield to pass too far east to affect many locations in the Mid-Atlantic and Northeast where heavy snow was observed. The NCEP SREF and experimental parallel SREF brought the storm and precipitation shield farther west, similar to that produced several days in advance by the European Centers model. This posed serious forecast issues due to the widely divergent forecasts associated with the quantitative precipitation shield.

The larger scale pattern as shown by the 500 hPa heights from11 to 16 February (Fig. 2) showed an emerging ridge with above-normal 500 hPa heights along the West Coast and southwestern United States and a wave moving over the western ridge (Fig. 2a-c) which deepened as it moved eastward. A second northern stream wave (Fig. 2d) over the Great Lakes at 0000 UTC 14 February interacted with the wave along the coast, which had a closed 5340 m contour over the Delmarva at 0000 UTC 14 February. This cut-off was associated with the enhanced precipitation from Virginia into southern New England (Fig. 1). The 12-hourly QPE data (Fig. 3) shows the enhanced area of precipitation over Maryland between 1200 13 February and 0000 UTC 14 February 2014 (Fig. 3c) and how this enhanced area of precipitation moved into New York and New England.

The storm evolved in a more complex fashion than most operational models implied. The interaction between the northern and southern stream waves created uncertainty issues as each model and ensemble forecast system likely handled this phasing differently. The problems associated with phasing wave merger cyclogenesis (WGC) have been document by Hakim et al (1995) and Gaza and Bosart (1990). When two waves are involved in cyclogenesis, complex storms can evolve and these WGC events may not be well-predicted by numerical forecasts.

This paper will examine the winter storm of 12-14 February 2014. The overall pattern is presented, with a focus on where the precipitation fell and the pattern in the context ofstandardized anomalies. Forecasts are presented focused on the uncertainty associated with forecasts of this storm. The focus is on the NCEP GEFS and SREF systems and will include forecasts for the experimental SREF. The GEFS, perhaps due to its coarser 55 km resolution, did poorly on this event relative to the 16km EC and 16km SREF.

2. 
   Methods and Data
   

Ensemble forecasts were derived from the NCEP Global Ensemble Forecast System (GEFS) and the Short Range Ensemble Forecast System (SREF). The surface and 500 hPa pattern were used to show how the general forecasts of a significant storm were predicted at longer ranges.

As with nearly all weather events, the high impact weather, including the heavy snow, sleet, and freezing rain occurred in relatively narrow corridors and was not as well predicted, even at longer ranges. For the snowfall, a 16 mm contour was chosen in the Probability of QPF images (POP) as it is close to 0.60 inches and using a first guess 10:1 snow ratio it would support heavy snow in central Pennsylvania. Climatologically, 12 to 13:1 is closer to the snow ratios in early February depending on location. Areas near the rain/snow and freezing rain area clearly did not meet the 10:1 criteria for snowfall.

Snowfall data was retrieved from the National Snow site in text format, decoded in Python, and plotted using GrADS. The QPE data was retrieved from the Stage-IV 6-hour data. These data too were plotted using GrADS.

The GEFS mean was used to illustrate differences in the intensity of the 500 hPa heights between different GEFS cycles. The pattern was defined as forecast minus observed and, when using two forecasts, most recent minus older forecast cycle. Thus positive (negative) values imply a stronger (weaker) analysis or more recent forecast relative to the older forecasts.

The European Center (EC) model was retrieved from the EC TIGGE site for post analysis. It should be noted EC forecasts for snow were readily available on social media days before the event.

3. 
   Pattern
   

At the surface (Fig. 5) a retreating cold anticyclone (Fig. 5a) was present over the northeast with a cyclone tracking across the northern Gulf of Mexico. A classic cold air damming (Forbes et al 1987; Richwien 1980) situation contributed to the snow and ice in the Deep South. The 850 hPa temperatures (Fig. 6) showed the cold air with areas of sub-freezing temperatures in the Gulf States. Note as the storm deepened and travelled up the coast, it pulled in warm air with +1 above normal temperature anomalies. This warm air caused snow to change to rain from Virginia to southern Connecticut.

The 850 hPa winds showed (Fig. 7) strong easterly winds with -3 to -5 u-wind anomalies north of the southern stream surface and 850 hPa cyclones. Stuart and Grumm (2006) showed that the strong u-winds are good indicators of strong winter storms and when the u-wind anomalies are in the -4 or lower range, the storms are typically both meteorologically and climatologically significant events. Note that around the time of the trough interactions and the closing off the southern stream 500 hPa wave (Fig. 4) the u-winds with -4 to -5 anomalies expanded.

As the southern stream wave cut-off and was interacting with the northern stream wave, the surface cyclone deepened. Off the Mid-Atlantic coast, the sea-level pressure anomalies fell to the minus -3 to -4 range. Deep cyclone and strong negative u-wind anomalies were shown by Root e al (2007) to be good predictors of major East Coast winter storms and precipitation events.

For good measure, the 250 hPa winds (Fig. 8) showed weak easterly flow at 250 hPa as the jet in the southern stream moved over the southeastern United States. The strong southerly flow produced a jet exit circulation in this region while the strong u-wind anomalies to the north and east defined the strong jet entrance region to the northeast. Like many significant winter storms, (Kocin and Uccellini 2004) this storm had coupled jets.

4. 
   Ensemble Forecasts of the Pattern
   

1. 
   European Center forecasts 16 km (EC)
   

The QPF shield showed some coastal precipitation, mainly snow, from forecasts issued at 0000 UTC 8 February. The forecast from 9 February, with a close-to-coast-track moved the precipitation shield westward and produced a large 25mm contour over much of the interior areas as far west as State College (the black dot in Figure 10). These robust forecasts garnered a lot of attention on social media. Subsequent forecasts actually showed lower QPF amounts as the phasing wave issue impacted the EC. The refined structures in the EC QPF fields clearly show the model attempting to adjust to the phasing waves and the cut-off 500 hPa low which the system was attempting to evolve (not shown).

The robust EC forecast of 9 February was tempered and changed as the phasing wave merger cyclogenesis issues affected the forecasts.

2. 
   Global Ensemble Forecast System 55 km (GEFS)
   

This slower westward trend and weaker cyclone at longer ranges provided lower QPF amounts to the west (Fig. 12). The GEFS never produced sufficient QPF for heavy snow over most of interior Maryland and Pennsylvania. The large spread in the 16mm contours did show that there was considerable uncertainty with QPF, and much of this was associated with the spread in the cyclone track in the model (not shown).

The QPF details show that the EC was a good 24 hours ahead of the GEFS in dealing with the closed off wave and intensified QPF over the Mid-Atlantic region. This feature is not well defined until the forecast issue at 1200 UTC on 12 February 2014 (Fig. 12f).

3. 
   Short-Range Ensemble Forecast System (SREF)
   

The SREF QPF fields (Fig. 14) showed a trend to bring the 16 mm contour farther west with time. The enhanced QPF with the wave phasing issues showed a nice shape and character by 0900 UTC 9 February (Fig. 14d), but still had high QPF far enough west at high probabilities (Fig. 15). The black dot near State College showed the SREF was unable to produce a high probability of heavy snow QPF over central Pennsylvania.

From a pattern perspective, the SREF predicted a strong easterly jet at 850 hPa (Fig. 16). Many signals for a significant storm in the coastal plain were relatively well predicted.

4. 
   Short-Range Ensemble Forecast System Parallel (SREFPARA)
   

The plume diagram for State College showed a significantly higher spread (Fig.20) in the SREFPARA with up to 1.75 inches of QPF to fall as snow and a mean of 0.82 while the SREF showed a smaller threat of heavy snow with a maximum of 1.37 inches and mean of 0.49 inches of QPF. A snow band moved into State College producing 8-10 inches locally.

5. 
   Summary
   

The EC model produced the earliest solution of a storm tracking along the coast, and forecasts issued on 9 February produced a significant area to be affected by 25mm or more of QPF (Fig. 10). Relative to observations, this was not a particularly skillful QPF. The higher amounts of precipitation were shifted to the east; however, the western extent of the QPF shield was generally farther west in EC forecasts relative to the GFS (not shown) and GEFS. The EC clearly picked up on the stronger cyclone close to the coast relative to the NCEP models and EFSs. This may be due both the higher resolution of the model and the asynchronous data assimilation methods employed.

The NCEP SREF produced a more westward QPF shield than the GEFS, and the SREFPARA indicated a potential for more QPF and snow farther west than the SREF. The SREFPARA also showed a larger spread in the QPF relative to the operational SREF. It is possible that the uncertainty was associated with the phasing waves an issue which likely impacted all the modeling and ensemble forecast systems.

This case seems to imply that the wave phasing issue and the development of the cut-off created considerable uncertainty in the forecasts. The coarser resolution models appeared to be slower to deal with the wave phasing issues and had more difficulty addressing it. This likely lead to the poorer GFS-GEFS QPF and cyclone tracks. The SREF did a bit better with this and the SREFPARA a bit better too. Clearly, the SREFPARA covered the spread better relative to the other two NCEP EFSs.

This case involved phasing waves and what proved to be a complex winter storm. There was considerable uncertainty with this storm. How to best forecast a storm in the face of conflicting guidance, even when one single model is generally viewed as superior, is a difficult chore. A poorman’s ensemble (PME) and a super-blend are likely the means to deal with events of this nature.

6. 
   Science Issues
   

• 
  Forecast system conflicts
  
  • 
    EC outperforms GFS and thus GEFS over long range verification scoring
    
  • 
    How should forecasters deal with model ambiguities and when the more significant storm is in the more skillful mode?
    
  • 
    A Poorman’s ensemble weighted toward the higher skill models based on verification?
    
  • 
    SREF and GEFS systems diverged with more robust QPF in SREF verse GEFS raises same EC/GFS issues. Though the EC QPF diminished after 9 February.
    
• 
  Ensemble use
  
  • 
    dealing with the edges of precipitation shields is a difficult issue.
    
  • 
    The SREFPARA showed larger spread in the QPFs which appeared to have been related to the phasing wave issue.
    
  • 
    How do forecasters who think deterministically when dealing with larger uncertainty?¹
    

7. 
   Acknowledgements
   

8. 
   References
   

Baker, B.W.1960: The 1960 ice storm in northern Alabama. Weatherwise, 13,196-200.

Bell, B. D. and L. F. Bosart, 1988: Appalachian cold-air damming. Mon. Wea. Rev., 116, 137–161.

Doty, B. E., and J. L. Kinter III, 1995: Geophysical data and visualization using GrADS. Visualization Techniques Space and Atmospheric Sciences, E. P. Szuszczewicz and Bredekamp, Eds., NASA, 209–219.

Forbes G. S., R. A. Anthes, and D. W. Thompson, 1987: Synoptic and mesoscale aspects of an Appalachian ice storm associated with cold-air damming. Mon. Wea. Rev., 115, 564–591.

Gaza, B. and L. F. Bosart, 1990: Trough merger characteristics over North America. Wea. Forecasting, 5:314–331. [Abstract]

Grumm, R.H. and R. Hart. 2001: Standardized Anomalies Applied to Significant Cold Season Weather Events: Preliminary Findings. Wea. and Fore., 16,736–754.

Gyakum J. R., and P. J. Roebber, 2001: The 1998 ice storm—Analysis of a planetary-scale event. Mon. Wea. Rev., 129, 2983–2997. Find this article online

Hakim, G. J., L. F. Bosart, and D. Keyser, 1995: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part I: Multiscale case study. Mon. Wea. Rev, 123:2663–2692

Hamill, T.M, 2003: Evaluating Forecaster’s Rule of Thumb: A study of d(prog)/dt. Wea. Forecasting,18,933-937.

Harlin, B.W. 1952: The great southern glaze storm of 1951. Weatherwise,5,10-13.

Hart, R. E., and R. H. Grumm, 2001: Using normalized climatological anomalies to rank synoptic scale events objectively. Mon. Wea. Rev., 129, 2426–2442.

Hoffman,R.N. and E. Kalnay, 1983: Lagged Average Forecasting. Tellus, 35A,100-118.

Kocin, P. J., and L. W. Uccellini, 2004: Northeast Snowstorms, Volume I: Overview. Meteor. Monogr., Vol. 32, No. 54, Amer. Meteor. Soc., 1-296.

Kocin, P. J., and L. W. Uccellini, 1990: Snowstorms along the northeastern Coast of the United States: 1955 to 1985. Meteor. Monogr., No. 44, Amer. Meteor. Soc., 280p.

Irland L. C., 2000: Ice storm 1998 and the forests of the Northeast. J. For., 96, 32–40. Find this article online

Rauber R. M., M. K. Ramamurthy, and A. Tokay, 1994: Synoptic and mesoscale structure of a severe freezing rain event: The St. Valentine's Day ice storm. Wea. Forecasting, 9, 183–208. Find this article online

Richwien, B. A., 1980: The damming effect of the southern Appalachians. Natl. Wea. Dig., 5(1), 2–12.

Robbins, C.C and J.V. Cortinas 2002: Local and synoptic environments associated with freezing rain in the contiguous United States. Wea. Forecasting,17,47-65.

Root, B., P. Knight, G.S. Young, S. Greybush, R.H. Grumm, R. Holmes, and J. Ross, 2007: A fingerprinting technique for major weather events. Journal of Applied Meteorology and Climatology, 46, 1053–1066.

Stewart R. E., 1992: Precipitation types in the transition region of winter storms. Bull. Amer. Meteor. Soc., 73, 287–296. Find this article online

Stewart R. E., and P. King, 1987: Freezing precipitation in winter storms. Mon. Wea. Rev., 115, 1270–1279. Find this article online.

Wandishin,M.S, M.E Baldwin, S.L. Mullen, and John V. Cortinas: 2005: Short-Range ensemble forecasts of precipitation type. Wea. And Forecasting,20,609-626.

Wetzel,S.W. and J.E. Martin 2001: An operational ingredients based methodology for forecasting midlatitude winter season precipitation. Wea. and Fore.,16,156-167.

Sivillo, S.K,J.E. Ahlquist, and Z. Toth,1997: An ensemble forecasting primer. Wea. Forecasting.,12, 809-818.

Stuart, N. A., R. H. Grumm, and M. J. Bodner, 2013: Analyzing predictability and communicating uncertainty: Lessons from the post-Groundhog Day 2009 storm and the March 2009 “megastorm.” J. Operational Meteor., 1 (16), 185–199.   

Stuart,N.A and R.H . Grumm 2006: Using Wind Anomalies to Forecast East Coast Winter Storms. Wea. and Forecasting, 21,952-968.

Zsoter, Ervin, Roberto Buizza, David Richardson, 2009: “Jumpiness” of the ECMWF and Met Office EPS Control and Ensemble-Mean Forecasts. Mon. Wea. Rev., 137, 3823–3836

Figure . Stage-IV estimated liquid equivalent precipitation from 000 UTC 12 through 1200 UTC 14 February 2014. Shading indicates values in millimeters. Contours are 16 and 50 mm. Return to text.

Figure .The GFS 00-hour forecasts of 500 hPa heights (m) and 500 hPa height anomalies in 24 hour increments from a) 0000 UTC 11 February 2014 through f) 0000 16 February 2014. Height contours every 60 m and standardized anomalies in standard deviations from normal. Return to text.

Figure . As in Figure 1 except over the Mid-Atlantic region showing 12-hour accumulated QPE for the 4 periods ending at a) 0000 UTC 13 February, b) 1200 UTC 13 February, c) 0000 UTC 14 February, and d) 1200 UTC 14 February 2014. Return to text.

Figure . As in Figure 2 except over the eastern United States and in 6-hour increments from a) 0000 UTC 13 February through f) 0600 UTC 14 February 2014. Return to text.

Figure . As in Figure 4 except for mean sea-level pressure and pressure anomalies. Return to text.

Figure . As in Figure 4 except for 850 hPa temperatures and temperature anomalies. Return to text.

Figure . As in Figure 6 except for 850 hPa winds and 850 hPa u-wind anomalies. Return to text.

Figure . As in Figure 7 except for 250 hPa winds. Return to text.

Figure . European Center mean sea-level pressure (hPa) forecasts form the EC model shown with standardized anomalies. Forecasts are valid at 1800 UTC 13 February 2014 from forecasts initialized at a) 0000 UTC 8 February, b) 0000 UTC 9 February, c) 0000 UTC 10 February, d) 0000 UTC 11 February, e) 0000 UTC 12 February, and f) 1200 UTC 12 February. Data courtesy of the EC TIGGE website. Return to text.

Figure . As in Figure 9 except for EC accumulated 24 hour QPF for the period ending at 0000 UTC 14 February 2014. Return to text.

Figure . As in Figure 9 except for the mean of the 21-member GEFS system and standardized anomalies. Return to text.

Figure . As in Figure 11 except for GEFS mean QPF (mm) and each members 16mm contour (thin colored lines) with a thick black contour showing the ensemble mean 16mm contour Return to text.

Figure . As in Figure 11 except for SREF mean pressure and anomalies from SREF initialized at a) 0900 10 Feb, b) 0900 11 Feb, c) 2100 UTC 11 Feb, d) 0900 UTC 12 Feb, e) 1500 UTC 12 Feb, and f) 2100 UTC 12 Feb 2014. Return to text.

\

Figure . As in Figure 13 except for SREF mean QPF (shaded) and each members 16mm contour and the mean 16mm contour in thick black. The black dot is State College, PA. Return to text.

Figure . As in Figure 14 except for the probability of 16kmm or more QPF. Return to text.

Figure . As in Figure 13 except 850 hPa winds and u-wind anomalies. Return to text.

Figure . As in Figure 13 except SREF PARA. Return to text.

Figure . As in Figure 14 except SREF-PARA. Return to text.