|Eastern United States Snow event of 5 February 2014
Dealing with uncertainty and varying predictability horizons
Richard H. Grumm
National Weather Service State College, PA 16803
A winter storm brought snow from the southern Plains to the Mid-Atlantic region on 4-5 February 2014 (Fig. 1). In the warm sector, heavy rainfall was observed over the Mid-Mississippi Valley (MMV) from eastern Arkansas into western Kentucky (Fig. 1a). The winter storm was associated with a strong 500 hPa trough (Fig. 2) and large ridge over the western Atlantic. The flow between the ridge and trough produce a surge of warm moist air into the eastern United States. In the MMV this led to heavy rainfall and in the Mid-Atlantic region, the warm air over retreating cold air produce a significant ice storm across southeastern Pennsylvania into New Jersey.
The pattern over the eastern United States, with the large ridge over the western Atlantic and the strong baroclinic zone north of the ridge is an established pattern often associated with ice storms (Baker 1960;Bell and Bosart 1988: Gyakum and Roebber 2001; Rauber et al. 1994; Wetzel and Martin 2001). A key component to predicting ice storm is where the intrusion of warm will occur and how enduring the low-level cold air persists during the precipitation event. Robbins and Cortinas (2002) examined the local scale environments with ices storms. Studies related to cold air damming (Richwien 1980;Forbes et al. 1987; Bell and Bosart 1988) demonstrated how the pattern and terrain features of the eastern United States can aid in keeping shallow cold air near the surface as warm air is present aloft.
Ice storms can be particularly damaging to trees (Irland 2000) and power lines. The weight of the ice causes trees and tree limbs to split and fall. Falling limbs and ice loading on power lines can cause power outages. New reports suggested that hundreds of thousands of customers lost power (AP 2014) during the ice storm of 4-5 November 2014. Widespread Power outages were observed across a wide swath of southern and southeastern Pennsylvania. Reports indicated hundreds of thousands of customers were without power for 1 or more days. Initial reports suggested this was one of the more devastating ice storms in southeastern Pennsylvania in over the last 10-20 years.
This paper will examine the pattern in which this winter storm developed and the predictability issues associated with this event. The pattern is addressed using the synoptic pattern and the standardized anomalies which when applied to forecasts of the pattern often aid in identifying high impact events. Forecasts are addressed from and NCEP ensemble prediction system perspective.
Methods and Data
The large scale pattern was reconstructed using 00-hour forecasts from the NCEP Global Forecast System (GFS). Standardized anomalies were computed as in Hart and Grumm (2001) using the re-analysis (Climate Forecast System; CFS) climate (R-Climate). The climatology spans a 30 year period. All data were displayed using GrADS (Doty and Kinter 1995).
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 will 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 were not as well predicted, event at longer ranges. For the snowfall 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) imply a stronger (weaker) analysis or more recent forecast relative to the older forecasts.
The evolution of the 500 hPa pattern (Fig. 2) showed the ridge over the western Atlantic with +1s 500 hPa height anomalies and a strong wave moving out of northwestern North America. As the distance between these two systems decreased, the southerly flow between them increased allowing for a surge of warm moist air into the southern United States. At the surface a wave moved along the western Gulf and into the MMV and Ohio River Valleys (Fig. 3). A secondary cyclone developed along the East Coast along a low-level baroclinic zone.
The secondary redevelopment was too far north, along the coast of Long Island, which allowed for a surge of warm air at both 700 and 850 hPa (Figs. 4 & 5). The strong ridge (Fig. 2) over the western Atlantic suggested a strong surge of warm air and the 700 hPa temperatures were above 0C over much of central and southern Pennsylvania and points south from around 0600 and before 1200 UTC 5 February 2014 (Fig. 4). At 850 hPa the wedge of cold air stayed well entrenched below the 700 hPa warm air until sometime after 0600 UTC and the 0C line at 850 hPa covered the southern 2/3rds of Pennsylvania by 1200 UTC after the temperatures at 850 hPa began to fall. A cold front pushed through the region between 1200 and 1800 UTC and the 850 hPa temperatures fell below 0C over from west to east. The short-lived warm surge was over the western Atlantic by 0000 UTC 6 February 2014.
The wedge of warm indicated sleet would likely mix with snow and the persistent 850 hPa cold, indicative of low-level cold air were good signals for an mixed precipitation and potential ice storm event. The 850 hPa winds and the track of the 850 hPa cyclone also indicated conditions favoring mixed precipitation over the Mid-Atlantic region. The cold conveyor of anomalously strong 850 hPa winds mainly moved of the Kansas and Missouri at 1800 UTC 4 February (Fig. 6), across the Midwest and into New York State over the ensuring 24 hours (Figs. 6b-f). The rain in the MMV and Ohio Valley (Fig. 2) and the mixed snow to ice in the Mid-Atlantic region were observed in the more southerly flow at 850 hPa (Fig. 6b-e).
The 250 hPa winds (Fig. 7) showed the strong jet and implied jet entrance region over the 500 hPa ridge (Fig.1) and implied 250 hPa ridge (not shown). The implied jet entrance region and the implied thermally direct circulation are well known signal to produce cold air damming and is a common characteristic associated with both ice storms and snow storms.
Ensemble Forecasts of the pattern
The NCEP GEFS showed considerable skill in predicting the evolution of this storm and the combined GEFS and CME EFS showed a potential winter storm from forecasts issued as early as 9 days before the event (Fig. 8). Some patterns are inherently more predictable than others.
Global Ensemble Forecast System (GEFS)
Forecasts from 6 NCEP GEFS cycles initialized at 0000 UTC 29 January through 3 February 2014 show that the GEFS predicted a cyclone (Fig. 9) along the coast at 1200 UTC 5 February and it predicted the strong 500 hPa trough and ridge pattern (Fig. 10). The strength and position of the 500 hPa ridge over the western Atlantic varied considerably from run-to-run with several runs from the 0000 and 1200 UTC (not shown) cycles bringing the 5880 m contour into or near the East Coast of Florida.
Forecasts from six 1200 UTC GEFS forecasts of the surface pattern (Fig. 11) show a similar pattern to that provided by the 0000 UTC members. Some earlier forecasts were colder and implied a coastal cyclone track which favored snow (see for example Fig. 11a). The forecast made at 1200 UTC 31 January showed a stronger cyclone farther west implying a warm solution. Though not shown, forecasts of precipitation type shifted from snowier to more mixed to more rain as the cyclone track shifted westward and then back to a more eastward track (not shown). The salient point was that with over 1 week of lead-time a potential winter storm over the eastern United States was predicted, though the details were not exact.
The GEFS quantitative precipitation forecasts (QPF: Fig. 12) from the 1200 UTC cycles showed a larger area to be impacted by QPF and 25mm or more of QPF. Several GEFS members also showed the potential in the 36 hour period for over 50 mm of QPF (Fig. 13). There was considerable cycle-to-cycle variability, with the salient point of a significant winter storm which could produce a larger swath of 25 to 50 mm of QPF. The highest threat for high QPF values was focused over the Ohio Valley (Fig. 13).
Shorter range GEFS showed a convergence of solutions. However the transition to the shorter-range is presented using the NCEP 16km SREF.
Short range Ensemble Forecast System (SREF)
Similar to the GEFS, the SREF predicted a strong winter storm and a secondary cyclone to develop along the coastal plain after the main surface cyclone tracked into western Pennsylvania (Fig. 14). North of the surface and 850 hPa cyclones, the SREF predicted the cold conveyor belt as depicted by the 850 hPa winds and wind anomalies (Fig. 15). The snow was forecast and fell in and near the warm conveyor, where 850 hPa temperatures (Fig.16) were predicted to stay at or below 0C. The snow to ice and the rain fell in the warm south to southwesterly flow.
The SREF precipitation type forecasts showed the snow, with higher probabilities in the cold air (Fig. 17) and ice in the area where the model generally predicted the shallow cold air. The SREF cold damming forecasts were generally useful though they under predicted the threat of ice in eastern and southeastern Pennsylvania with the higher probabilities focused over the higher terrain of western Maryland and southwestern Pennsylvania. Shorter range SREF freezing rain forecasts slowly increased the threat to the east (Fig. 18).
The SREF freezing rain forecasts are shorter in duration as the SREF emphasized the freezing rain farther west and the predictability horizon in southeastern and eastern Pennsylvania was quite short.
The SREF QPF showing the mean (Fig. 19) and the 25 mm contours (Fig. 20) and the probability of 25 mm or more QPF showed a similar, though more detailed evolution than that presented by the GEFS. Note the 12.5mm contour over eastern Kansas across Missouri with the snow on the north side of the evolving cyclone and the extent of the 25mm contour across the MMV and Ohio Valleys.
The 2100 UTC 4 February plume diagrams show the rapid rate of accumulation of the precipitation at Allentown and Harrisburg (Fig.21). These data the point values and the potential for significant ice accretion at both locations. The 2m temperatures (Fig. 22) showed the temperatures going above 32F at Harrisburg as the precipitation was winding down and the above freezing values near 850 hPa where temperatures peaked near 4C (Fig. 22). The warm air was present at Allentown though the 2m temperatures remained below 0C.
A significant winter storm brought snow from eastern Kansas and Missouri, across the Midwest into northern Pennsylvania and New York State on 4-5 February 2014. The overall pattern and the potential for a winter storm in the eastern United States was relatively well predicted with at least 8 days of lead-time in both the NCEP GEFS and US-Canadian 42 member NAEFS. The details as to where the heavier snow and higher precipitation amounts were not as well predicted as the overall pattern. The predictability horizon of the storm and region impacted was on the order of days. The details of the areas for heavy rains, heavy snow, and ice were not as well predicted. The ice forecasts in southeastern Pennsylvania had high confidence predictability horizons on the order of tens of hours.
The pattern in which this storm developed is a pattern common with many winter storms and ice storms. The strong ridge of the western Atlantic played a critical role in the winter storm of 4-5 February 2014. The strong gradient between the trough and ridge (Fig. 1) allowed for a surge of warm air and above freezing temperatures as far north as the Mid-Atlantic region (Fig.4-5). The strong gradient on the northern edge of the ridge produced a strong jet streak (Fig. 7) and a thermally direct circulation which helped maintain the baroclinic zone and maintain the low-level cold air. Nearly ideal conditions for a winter storm with sub-freezing air maintained at the surface and a wedge of warm air in the layers above the surface from at least 850 to 700 hPa.
The heavy snow was along the northern edge of the strong southerly flow (Fig. 6). Snow and heavy snow were also observed in the strong easterly flow north and west of the track of the 850 hPa. The anomalously strong 850 hPa winds indicated a strong cold conveyor in the cold air. These features implied a strong jet entrance region (Fig. 7) which is often present during winter storms and particularly in ice storms.
The over GEFS forecasts correctly predicted the potential for both the storm and high QPF amounts with 7-8 days of lead-time. These forecasts of the storm and the Miller-B evolution were rather impressive. However the details remained elusive and there were at times signals for an East Coast Snowstorm which did not occur. Social media users employed forecasts from single models and began advertising the potential for a big snow storm over the Mid-Atlantic region 5-7 days in advance of this storm which ended up being a messy event with little snow south of the Mason-Dixon Line.
The shorter term SREF forecasts did well with many aspects of the event, though the details with the rain-snow and freezing rain areas were slow to focus in on the higher threat in southeastern Pennsylvania. The SREF freezing rain forecasts (Fig. 18) are shorter in duration as the SREF emphasized the freezing rain farther west and the predictability horizon in southeastern and eastern Pennsylvania was quite short. In tight gradients, near boundaries, and in cold air damming situations, the details are difficult to ferret out and require vigilance. Despite how far in advance this event was predicted, the high impact weather, to include the devastating ice was not well predicted with more than 1-2 days of lead-time in eastern Pennsylvania.
These data show that despite successful long-range forecasts, the important details are not as easily ferreted out.
This case study reveals a number of science issues the must be addressed before more accurate forecasts of these types of events are possible.
the larger scale pattern and threat for a storm had 7-8 days of predictability
the precipitation type issues were not as well predicted
details on the ice were hard to get with high confidence more than 6-24 hours in advance
Impacts for users
the ice had severe impact people with power outages and travel related issues.
The impact of the internet and social media
early forecasts of snow and a more coastal track caused social media alerts for a major snowstorm in the Mid-Atlantic region and northeast.
these forecasts made no visible use of ensemble data and were not necessarily retracted effectively.
The Pennsylvania State University Department of Meteorology for data access and discussions related to this storm.
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.
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.
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
Harlin, B.W. 1952: The great southern glaze storm of 1951. Weatherwise,5,10-13.
Irland L. C., 2000: Ice storm 1998 and the forests of the Northeast. J. For., 96, 32–40. 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.
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
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.
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.
Grumm, R.H. and R. Hart. 2001: Standardized Anomalies Applied to Significant Cold Season Weather Events: Preliminary Findings. Wea. and Fore., 16,736–754.
Hamill, T.M, 2003: Evaluating Forecaster’s Rule of Thumb: A study of d(prog)/dt. Wea. Forecasting,18,933-937
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.
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