Event Summary
     National Weather Service, Raleigh NC

Decmber 2-3, 2000 Winter Storm

  • Description - Two areas of snow were deposited in North Carolina during the period of December 2 - 4, 2000. Snowfall in the North Carolina mountains was associated with an upper air disturbance. The heaviest amounts (up to 9 inches) fell in the Southern Mountains, ahead of the track of a vorticity maximum embedded in a southeastward moving 500 MB short-wave. A more extensive area of snow occurred in the coastal plain and coastal areas in association with the development of surface low pressure in the coastal waters of the Carolinas. The largest snowfall totals (around a foot) were located in an area where weather radars showed persistent quasi-stationary snowbands. The transition zone from nearly all snow to rain along the immediate coast was quite narrow. Heavy snowfall (8 inches or more) in Eastern North Carolina is climatologically infrequent and is unusual, especially for early December.

    National Weather Service (NWS) forecast offices issued winter storm watches and warnings well in advance for those areas impacted by the snow. Watches and warnings (for 4 inches of snow or more) were also issued for those portions of North Carolina where little or no snow fell. Note the lack of snow across large portions of western and central North Carolina. This preliminary case study explores the forecast process used by the NWS Raleigh (RAH) and describes how this event unfolded.

    This case study will cover:
    – Early forecast thoughts
    – Heavy snow climatology: how unusual was this event?
    – Event evolution
        • Role of pre-storm precipitation on Saturday 12/2/00
        • Role of offshore baroclinic zone and coastal low
        • Role of incoming vorticity max
    – Model performance and model trends (dprog/dt)
    – Conclusion, and what we can learn from this event
        • QPF conceptual model
        • Performance of the TRENDS technique in forecasting of precipitation type
        • Summary and final thoughts

    Comments and suggestions are welcomed.

    Total snowfall 12/2/00 through 12/4/00
    Amounts are in inches and "T" represents trace amounts.

    Visible Satellite image showing snow cover on the ground at 1545Z on 12/4/00

  • Early Forecast Thoughts
    The philosophy behind winter weather forecasting at RAH has always been to stress accuracy with reasonable lead times over being the "first" to mention wintery precipitation, and we strive for a lead time of at least 12 to 18 hours for winter storm warnings. Model forecasts from 1200 UTC Saturday 12/2 continued to indicate (as they had done for several runs) a major precipitation event for North Carolina as the strong vorticity max approached from the west and the surface low intensified just off North Carolina's coast. Precipitation-type forecasting tools such as the TRENDS technique (see below) pointed to a mostly snow event for the RAH county warning area (CWA), with a little sleet mixed in over the coastal plain and eastern sandhills. With the event expected to begin Saturday night and continue through Sunday, and with models agreeing on snowfall amounts well above the warning criteria, this forecast shift decided on the issuance of a winter storm warning for the entire CWA.

    Eta MSLP and QPF forecast from 12Z Saturday 12/2/00, valid 18Z Sunday 12/3/00

  • Heavy Snowfall Climatology in the Carolinas - How unusual was this event?
    The probability of a heavy snowfall (8 inches or greater) in any given winter season is shown. East of the mountains, these events are climatologically relatively infrequent; however, when considering the record breaking snow from the January 24, 2000 explosive coastal low event, some counties in Eastern North Carolina have now received two heavy snowfall events in less than one year.

    Probability of a heavy snowfall (8 inches or greater) in any given winter season.

    Average seasonal snowfall, in inches, across North Carolina.

  • Evolution of the Heavy Snow Event- Water vapor imagery shows the important upper level features. As the imagery loops, note that there are two disturbances in the northwest flow upstream from North Carolina. A loop of radar imagery shows an area of rain moving eastward from central North Carolina during the day on December 2, 2000. This rain was associated with the lead upper level disturbance and may have influenced the distribution of snow that followed in eastern North Carolina. The possible effects of the lead vorticity disturbance included pushing the coastal baroclinic zone seaward, initializing the development of the coastal low sooner, and precipitating what little moisture was over central North Carolina prior to the arrival of the more vigorous vorticity disturbance. All of these effects would potentially limit the occurrence of snow in central NC and shift the heavier snow farther eastward than might have otherwise been the case.

    The magnitude and depth of the surface based cold air over North Carolina is not precisely known since the RAOB at Greensboro NC (KGSO) was not available. In fact on December 2, 2000 at 12Z, a number of RAOB's within or near the cold-air damming region were missing and potentially may have affected the model's solution regarding the strength of the damming wedge. It is reasonable to conclude that the strength of wedge did in fact limit the occurrence of precipitation in the snow deficit areas of western and central North Carolina.

    The water vapor imagery also revealed additional information. Note the decreasing definition in the vorticity maximum as it crosses the NC mountains. This may be signaling a decrease in the system's potential for lift. Indeed, the radar imagery shows the area of snowfall diminishing as it exits the mountains into the Upstate of South Carolina. Not only did this weakening lead to lesser precipitation amounts just east of the mountains, but it also had implications on the potential for energy transfer from the upper system to the coastal surface low. Energy transfer can be viewed as linkage between the lift generated by the upper level dynamics to the converging low level inflow associated with the coastal low. The potential for linkage between the two systems diminishes as the lift with the upper system lessens and/or as the convergent moist low-level inflow with the coastal low decreases. As the potential for energy transfers lessens so does the potential for precipitation, especially in central North Carolina (damming region) where system interaction is most critical.

    Relative to the observed low-level converging moist inflow, there are reasons to believe that the model's forecast of inflow was overdone. Note the ETA's forecast for the 850 MB flow associated with the coastal system. A 40 to 50 kt east-northeasterly inflow was predicted at 12Z December 3, 2000. But real-time analysis of VAD winds as the event unfolded found the 850 flow at RAH to be much less (20 kts) and showing a more northerly component. This too limits the potential for precipitation due to less convergence, less moisture, and less interaction between the two systems. The lack of a moisture laden upper level jet also likely limited the development of the coastal low.

    These noted inferences were used by RAH forecasters to trend down the initial forecast snowfall amounts in the western and central portions of the RAH CWA. A regional radar image from 0630Z December 3, 2000 shows the pattern of precipitation coverage as the event was beginning to unfold. A slow backing of the precipitation from the coast into eastern NC gave confidence to the forecast of heavy snow in the coastal plain. Note the weaker precipitation returns seen in central NC between the two main areas of precipitation. The enhanced precipitation elements seen within the two main areas were thought to represent convection that could possibly enhance and help to fill in the precipitation gap in central NC. While confidence in potential interaction and energy transfer between the two systems was diminishing, the potential for latent heat release and its effects on the development of precipitation was enough to keep RAH forecasters committed to a significant snow event in central NC. Warnings in central NC were later dropped in an early morning update on December 3, 2000.

    Water Vapor Image valid 0915Z Sunday 12/3/00

    Loop of Water Vapor Imagery from 1815Z 12/1/00 through 2315Z 12/3/00

    Regional reflectivity image valid 0630Z Sunday 12/3/00

    Regional reflectivity loop from 1800Z 12/2/00 through 0800Z 12/4/00

    KRAX base reflectivity image valid 1725Z Sunday 12/3/00

    KRAX base reflectivity loop from 0004Z 12/3/00 through 0304Z 12/4/00

  • Surface Analyses -
    09Z 12/3/00 Surface Analysis

    12Z 12/3/00 Surface Analysis

  • Model Performance - This event was associated with three principal meteorological features - a vigorous vorticity maximum embedded in a northwest flow 500 MB short-wave, a strong cold-air damming wedge, and a developing surface coastal low. The models generated significant QPF across much of North Carolina as the short-wave crossed the North Carolina mountains, while a coastal surface low developed off the Carolinas. The models under forecast a lead 500 MB vorticity disturbance, which was located east of the vigorous vorticity maximum. As the lead vorticity disturbance moved across NC on Saturday December2, it produced measurable rain over a large area of the state. This feature may have affected the pattern of QPF (snow) that occurred across North Carolina late on December 2 through early December 4, 2000.

    In general the models captured the overall synoptic pattern for the event. The models QPF fields showed significant errors in both the magnitude and placement of the precipitation. Relative to the ETA, there was less error in the AVN's QPF fields. The AVN was more consistent in showing an open short wave at 500 MB thus less intense development of the surface coastal low. As the winter storm warnings were being issued, both odels portrayed significant amounts of precipitation in central North Carolina where little occurred. Although the ETA did trend toward a more accurate QPF prediction, the trend was not apparent to RAH forecasters until after the issuance of the warnings. In addition, the ETA consistently forecasted significant QPF amounts which were hard for forecasters to ignore.

    Raw Model Output - Raw output from the ETA Fous data for Raleigh-Durham is shown below.

    The ETA 6 hourly QPF amounts are shown in column format with the 48 hour model total QPF depicted. Note the models consistently depicted QPF amounts near or in excess of 1.0 inches at KRDU for seven consecutive runs. Also note that the 12Z model run on 12/3/00 still forecast 0.28 inches at KRDU and the 18Z 12/3/00 model run forecast 0.06 inches for the first 6 hourly period while some sunshine was breaking out at KRDU.

    Only a trace (T) of snow was observed at Raleigh-Durham (KRDU).

    ETA Fous data
    12/01/00 12/02/00 12/0200 12/0200 12/0200 12/03/00 12/03/00 12/03/00 12/03/00
    18Z 00Z 06Z 12Z 18Z 00Z 06Z 12Z 18Z
    Fcst Hour 00 0 0 0 0.04 0.11 0.06 0.09 0.11 0.06
    Fcst Hour 12 0 0 0 0.14 0.25 0.24 0.26 0.17 0
    Fcst Hour 18 0 0 0.12 0.27 0.44 0.21 0.46 0 0
    Fcst Hour 24 0 0 0.37 0.38 0.53 0.73 0.16 0 0
    Fcst Hour 30 0.01 0.07 0.51 0.63 0.53 0.22 0 0 0
    Fcst Hour 36 0.20 0.31 0.47 0.74 0.53 0 0 0 0
    Fcst Hour 42 0.29 0.80 0.59 0.54 0 0 0 0 0
    Fcst Hour 48 0.45 0.81 0.31 0.03 0 0 0 0 0
    Total 0.95 1.99 2.38 2.77 2.39 1.46 0.97 0.28 0.06

  • Model Trends (dProg/dt) - Trends in model solutions may show a more accurate solution than a given specific model run. NCEP demonstrated this to be the case with the record breaking snowfall on 24 January 00. Model trends--known as Dprog/DT procedures on AWIPS--show different model runs valid at the same time. AWIPS allows forecasters to compare 3 Eta model runs. At the time NWS offices issued winter storm warnings, there was little in the way of significant trends indicated in the ETA’s various forecast fields. But after the warnings were in place, the trends by the ETA were to decrease QPF amounts and shift the QPF pattern and the location of the surface low southeastward. Though these trends proved to be valid, they were apparent to RAH forecasters only after the warnings had been issued.

    Below are forecast plots from successive Eta model runs, all valid at 18Z Sunday 12/3/00. Plots shown are (top to bottom) the 30-hr forecast, the 18-hr forecast, and the 6-hr forecast.

    • Mean sea level pressure (white), model QPF (black), and surface dewpoints (image) -

    30 hour forecast from 12Z 12/2/00 valid 18Z Sunday 12/3/00
    (Model run available to RAH forecasters when winter storm warning was issued)

    18 hour forecast from 00Z 12/3/00 valid 18Z Sunday 12/3/00

    6 hour forecast from 12Z 12/3/00 valid 18Z Sunday 12/3/00

  • QPF Conceptual Model
    This heavy snow event involved three primary meteorological features: a vigorous vorticity maximum embedded in a northwest flow 500 MB short-wave, a strong cold-air damming wedge, and the development of a surface coastal low. Previous winter weather events allow for some conceptualizations about the distribution of precipitation associated with this pattern of cyclogenesis. With this pattern, the precipitation distribution in the North Carolina mountains is dependent upon the track and intensity of the vorticity maximum and 500 MB short-wave. Precipitation distribution in eastern North Carolina is dependent on the location, track, and intensity of the coastal surface low. The highest degree of precipitation uncertainty is found in central North Carolina. Precipitation in this area is in large part dependent upon energy linkage between the upper level dynamics and the low-level baroclincity. Without energy transfer between the two systems there may well be a deficit of precipitation in central North Carolina as was the case in this event. Also, cold-air damming in the area can be another precipitation limiting factor given the depth of and degree of dryness in the surface based cold air.

  • Evaluating Precipitation Types -

    RAH forecasters used a number of forecast tools and concepts to correctly anticipate snow as the predominant wintry precipitation type. The station's conceptual models show that sufficiently well organized surface coastal lows are typically associated with a narrow transition zone of mixed precipitation. Such lows often produce a relatively deep near freezing isothermal layer when coupled with a dry and cold airmass already in place over inland areas. Hence, there is little icing found with this synoptic pattern outside of the narrow transition zone from rain to snow.

    The ETA's forecast soundings from the 00Z 12/3/00 model run are shown below for Rocky Mount (RWI). Around 7 inches of snow fell in RWI from about 11z-23z on 12/3/00. The p-type forecasts from the AREA technique which uses positive (geometric portion of sounding above freezing) and negative (geometric portion of sounding below freezing) areas to predict p-types are shown just to the left of the soundings. Note the forecast of precipitation types for the 11Z - 23Z period. While the precipitation is forecast to begin as snow, the technique quickly changes its p-type to several hours of freezing rain and later to a few hours of sleet. The AREAS's errors in its p-type forecast calling for a mostly icing event where 7 inches of snow occurred are directly due to the errors found in the details of the ETA's forecast soundings.

    The p-type forecast using RAH's TRENDs technique (based upon partial thickness) predicts measurable snow as the predominant p-type but likely mixing with a little sleet, rain, or freezing rain during the event. Note on the TRENDS's nomogram the appearance of the so-called "snowy nose" (the geometric triangle area). The snowy nose appears whenever there is a relatively deep near freezing isothermal layer on the sounding. The "snowy nose" is a prompt to the forecaster to expect measurable snow as the predominant p-type.

    ETA Forecast sounding for Rocky Mount, NC (KRWI) from 12/3/00 00Z ETA valid 12/3 06Z - 12/4 00Z

    Note - The red line is the forecast ambient temperature and the green line is the dew point temperature The light gray line is the omega field with negative values (to the left) representing upward vertical motion. The vertical axis is in thousands of feet.

    - Metars (observations) from KRWI for the same period.  The snow began at RWI at 1120Z on 12/3/00 and ended at 2255Z 12/3/00.

  • Summary and Final Thoughts - Though the NCEP models did a reasonable job of capturing the overall synoptic scale pattern for this event, the errors in the details were enough to provide guidance that pointed toward an inaccurate expectation of significant snow in central North Carolina and western North Carolina, east of the mountains. The errors were most prominent in the models QPF solution and especially so with the ETA's forecast. While trends in successive model runs did point to a more accurate QPF solution by the ETA, these trends were not apparent to RAH forecasters until after winter storm warnings had been issued.

    A conceptual model based upon case studies provides a helpful frame of reference from which to interpret model performance. Conceptual models also provide a means of interpreting the likely factors accounting for the distribution of precipitation with this and other similar events. While a number of factors account for the resulting precipitation pattern, energy transfer from upper level systems crossing the mountains, with coastal systems, appears to be critical for accurate QPF forecasts for central and western North Carolina, east of the mountains. A thorough understanding and familiarity with conceptual models such as the one discussed here, can only increase the forecaster's skills in correctly evaluating model guidance.

    The forecasting of precipitation type for this event was relatively straight forward. As the watches and warnings were being issued, RAH forecasters used a combination of the TREND technique and their knowledge of p-type distributions as a function of the pattern of cyclogenesis to correctly anticipate a snow event in eastern North Carolina. Errors in the details of the ETA's forecast soundings lead to the AREA technique projecting more of an icing than snow event in eastern NC.

      The following table summarizes the important factors and their impact on the distribution of precipitation across North Carolina with this event:

    Factors Effects

    Initial vorticity disturbance and the rain it produced

    Pushed the coastal baroclinic zone seaward; drained central NC of the limited moisture that had been in place; initiated earlier development of the low. All of which limited the westward progression of the precipitation shield.

    Strong cold-air damming wedge

    Prevented precipitation in the wedge area due to the magnitude and depth of cold air in place and created an less favorable environment for the transfer of energy between the two systems that night have lead to more precipitation in central North Carolina.

    Weakening of vigorous vorticity max as it crossed NC mountains

    Directly diminished the scope and magnitude of precipitation in the mountains and created an environment less favorable for the transfer of energy between the two systems that might have lead to more precipitation in central NC.

    Detection of weaker inflow from coastal system into central North Carolina

    Deprived very dry central NC of much needed moisture and limited the potential for the transfer of energy that might have lead to more precipitation in central North Carolina.

    Absence of a moisture laden subtropical jet

    Limited the development of the coastal low which in turn limited the westward extent of the precipitation and decrease the likelihood of energy transfer between the two systems that might have lead to more precipitation in central North Carolina.

    Lack of convection

    Limited the release of latent heat its potential for enhancing the development of precipitation.

  • Case study team -
    Phil Badgett
    Shaun Baines
    Jonathan Blaes
    Gail Hartfield
    Kermit Keeter

    For questions regarding the web site, please contact Jonathan Blaes.

  • NWS Disclaimer.