After heavy snow fell overnight, the sky cleared before sunrise on the morning of 2 March, providing for spectacular scenes of new snow clinging to trees under a bright blue sky such as this in Lincolnton, North Carolina.

Author's Note: The following report has not been subjected to the scientific peer review process.

1.  Introduction

Rain and snow occurred across the western Carolinas and extreme 
northeast Georgia on Sunday, 1 March, and Monday, 2 March 2009, while 
a low pressure system moved across the southeastern United States.  The 
precipitation began early on 1 March as rain at most locations, but snow 
occurred in the high elevations of the western Carolinas.  The rain 
gradually changed to snow over a larger area during the afternoon and 
evening as the snow level lowered.  By the time the precipitation ended 
during the early morning hours of 2 March, snow accumulations across the 
county warning area (CWA) of the Greenville-Spartanburg (GSP) Weather
Forecast Office (WFO) ranged from a trace to 16 inches (Fig. 1).  The 
most significant snowfall totals were in the high elevations east and 
northeast of Asheville, North Carolina, and in the Piedmont of both 
Carolinas.  Amounts less than one inch were observed in the far western 
mountain counties and in the northern foothills of North Carolina.  The 
higher snow accumulations extended northeast across the northern mountains 
of North Carolina and into the New River Valley of Virginia, while another 
swath of higher snowfall totals extended from the Piedmont Triad to south 
central Virginia.

Figure 1.  Storm total snowfall for the period 1-2 March 2009. Click on 
image to enlarge.

Click here for a list of snowfall reports for 1-2 March 2009

One of the unique features of this event was the simultaneous occurrence 
of snow and thunder during the evening of 1 March.  The axis of significant 
snow accumulation seen in Figure 1 extending from Upstate South Carolina 
into the Piedmont of North Carolina marked the area where snowfall rates 
were enhanced by lightning-producing convective storms.

2.  Synoptic Overview

The sequence of events that set the stage for the winter storm can be 
traced to 27 and 28 February when a cold front crossed the Appalachian 
Mountains (Fig. 2).  The front extended from the North Carolina coast 
southwest to the Texas coast at 1200 UTC on 28 February, as seen on the
surface analysis from the Hydrometeorological Prediction Center (HPC).
A cold high pressure air mass moved from central Canada into the northern 
Great Lakes and Great Plains.  During 28 February, a weak low pressure 
system developed on the front over the Gulf Coast states and moved 
eastward.

Click here to view a 21 frame Java loop of Surface Fronts and Pressure 
from 0000 UTC 28 February to 12000 UTC 2 March 2009.

Figure 2.  HPC surface fronts and pressure analysis at 0600 UTC on 
1 March 2009.  Click on image to enlarge.

The 500 mb height and wind fields at 1200 UTC on 28 February (Fig. 3) 
showed a ridge over the western United States and a trough over the 
Great Plains.  A 5460 m cutoff low over Kansas and Nebraska was the 
dominant feature of the trough.  The flow along the East Coast was 
generally from the west and southwest with a short wave trough moving 
into New England.  The flow was confluent over the northern mid-Atlantic 
region where the southwest winds merged with a northwesterly air stream 
circulating around a deep low over northeastern Canada.  The weak surface 
low on the cold front was over Georgia at 1800 UTC 28 February.  A 
surface trough extended from the low over Georgia northwest to Missouri 
and Kansas where the upper low was located.

Click here to view a 28 frame Java loop of GOES-12 Water Vapor Satellite
Imagery from 2345 UTC 27 February to 0345 UTC 1 March 2009.

Figure 3.  Objective analysis of 500 mb geopotential height, temperature, 
and wind barbs at 1200 UTC on 28 February 2009.  Click on image to 
enlarge.

At 0000 UTC on 1 March, the Rapid Update Cycle (RUC) model (80-km grid 
spacing) placed a 5410 m cutoff low at 500 mb over Arkansas (Fig. 4).  
Low 1000-500 mb thickness values (5210 m) accompanied the deep cyclonic 
circulation.  A wind maximum at 500 mb curved around the base of the 
trough with speeds in excess of 90 kt over northeast Texas.  The wind 
maximum at 300 mb over north central Texas was 125 kt.  The jet axis at 
300 mb extended northeast to the mid-Atlantic coast with 90 kt wind 
maxima over east Tennessee and Virginia.  The 500-300 mb omega analysis 
showed the most prominent area of upward motion was over west Tennessee.

Figure 4.  Objective analysis of 500 mb geopotential height (solid green 
contours, dm), 1000-500 mb thickness (dashed yellow contours, m) and wind 
speed (color fill, kt) at 500 mb from the RUC model at 0000 UTC on 1 March.  
The color table at the top shows the wind speed, shaded purple from 0 to 
20 kt and green and yellow greater than 65 kt. Click on image to enlarge.

The surface analysis at 0000 UTC on 1 March showed that a center of low 
pressure had moved to South Carolina with very little strengthening 
during the preceding 12 hours.  A weak surface trough extended from the 
low west to Arkansas where the upper-level cyclonic circulation was 
located.  The surface analysis at 0600 UTC on 1 March indicated the weak 
surface low was still over South Carolina, but another weak low was evident 
on the front in southern Georgia.  By 1200 UTC, a single surface low 
pressure system was near Savannah and the 5460 m cutoff low was over 
southern Alabama.  The 1003 mb surface low had deepened approximately 5 mb 
since 0000 UTC and the low was more organized in response to the approach 
of the upper-level trough, while the surface trough that extended west 
northwest to northern Alabama displayed better definition.

The 95 kt wind maximum at 500 mb had rounded the bottom of the trough 
at 1200 UTC and extended from the northeastern Gulf of Mexico into 
Georgia.  The 110 kt wind maximum at 300 mb was a little further west 
near the Louisiana coast, but the jet axis extended northeast into 
Georgia.  The Carolinas and Georgia were in the exit region of the wind 
maximum approaching from the Gulf of Mexico and in the entrance region 
of a 100 kt wind maximum over the Mid-Atlantic States.  The highly 
ageostrophic flow at 300 mb (Fig. 5) was accompanied by a 500-300 mb 
omega maximum over Upstate South Carolina and along the Carolina coast.

Figure 5.  Objective analysis of 300 mb isotachs (solid blue), ageostrophic 
wind (yellow barbs), and 500-300 mb omega (color fill) at 1200 UTC on 
1 March from the RUC-80 model.  Click on image to enlarge.

Between 1200 UTC 1 March and 0000 UTC 2 March, the surface low moved from 
coastal Georgia to the Outer Banks, and another surface low formed just 
inland over the Carolinas (Fig. 6).  The second low formed in the trough 
extending west from the coastal low and was a result of falling surface 
pressure associated with the deep and cold cyclonic circulation aloft that 
had been approaching the frontal boundary and primary low.  As the 500 mb 
low moved from southern Alabama to central South Carolina by 0000 UTC, the 
1000-500 mb thicknesses accompanying the cutoff gradually lowered across 
the area.  As a result of the lowering thicknesses, sufficient cooling 
occurred to produce snow at locations that had experienced rain during 
most of the day.

Figure 6.  HPC regional surface analyses at 1200 UTC on 1 March (upper
left), 1800 UTC 1 March (upper right), 2100 UTC 1 March (lower left), 
and 0000 UTC 2 March (lower right).  Click on images to enlarge.

During the next 12 hours the cutoff low at 500 mb evolved to an open, 
sharp trough that moved offshore.  The surface low moved to a position 
off the northern Mid-Atlantic Coast by 1200 UTC on 2 March.  The 
circulation around the offshore low and the high pressure system 
centered near the Great Lakes produced a dry northerly wind flow 
across the western Carolinas and extreme northeast Georgia.

Click here to view a 29 frame Java loop of a Regional Radar Reflectivity 
Mosaic from 0500 UTC 1 March to 0858 UTC 2 March 2009.

Click here to view a 22 frame Java loop of GOES-12 Water Vapor Satellite 
Imagery from 0645 UTC on 1 March to 0345 UTC on 2 March 2009.

Click here to view a 22 frame Java loop of GOES-12 Infrared Satellite 
Imagery from 0645 UTC on 1 March to 0345 UTC on 2 March 2009.

3.  Event Details

a. Sunday morning

Around 0530 UTC on 1 March, precipitation spread into the CWA from the 
southwest.  Rain was the dominant precipitation type, but subfreezing 
temperatures above approximately 6,000 ft MSL contributed to snow on 
the highest mountains.  Precipitation continued to increase in coverage 
so that by 1200 UTC light rain was beginning at Asheville, and rain was 
occurring over nearly the entire area east of the mountains. 

The precipitation developed as isentropic lift strengthened in advance 
of the approaching low, as seen on the 295 K surface obtained from the 
RUC13 model (the Rapid Update Cycle model with a 13 km horizontal grid
spacing) at 1200 UTC and the regional composite reflectivity radar 
mosaic (Fig. 7).  The wind field across Upstate South Carolina and the 
Piedmont of North Carolina showed the flow ascending the 295 K surface 
(from high pressure near the surface to low pressure aloft).  The 
precipitation coverage was quite similar to the area experiencing upward 
motion on that surface.

Figure 7.  Objective analysis from the RUC-13 model of pressure (green
 contours) and wind (yellow barbs) on the 295 K surface, with composite 
radar reflectivity mosaic (color fill) at 1200 UTC on 1 March.  Click 
on image to enlarge.

The 1200 UTC surface observations showed temperatures near 32o F in the 
extreme northeast portion of the CWA and below freezing in the higher 
elevations (Fig. 8).  Warmer temperatures (in the 40s) existed in the 
far western mountain counties because high terrain blocked the westward 
spread of the colder surface air over the Piedmont.  The observations 
indicated that most of the stations reporting precipitation experienced 
rain.

Figure 8.  Surface observations plot at 1200 UTC on 1 March.  Click on 
image to enlarge.

The RUC13 initial hour analysis thermal profiles at 1200 UTC on 1 March 
for four locations [Asheville (AVL), Charlotte (CLT), Hickory (HKY), 
and Greer (GSP)] showed that even though each location had surface 
temperatures only a few degrees above 32o F, a warm layer aloft 
prevented snow from reaching the ground (Fig. 9).  The horizontal red 
line on the AVL temperature profile indicates the height of Mt. Mitchell 
(6,684 ft MSL).  The near freezing temperature at that elevation and the 
subfreezing temperature above indicate that precipitation spreading 
across the mountains was probably in the form of snow at the highest 
elevations, but melting was occurring rapidly below 6,000 ft MSL.

Figure 9.  RUC13 initial hour profiles of temperature (red), dew point
(green), and wind (barbs) at 1200 UTC on 1 March for Asheville (AVL,
upper left), Charlotte (CLT, upper right), Hickory (HKY, lower left), 
and Greer (GSP, lower right).  The horizontal red line on the AVL 
profile indicates the height of Mt. Mitchell (6,684 ft MSL).  A wind
hodograph is shown in the upper left-hand corner of each image.  Click
on images to enlarge.

The eastward progression of the upper low center was expected to
strengthen the isentropic lift to the east of the center and the low
level cold advection around the western side as it acquired a more
negative tilt.  During the late morning hours, the strengthening of a
deformation zone to the north and west of the low allowed for the
development of a band of snow across north Alabama and north and west
central Georgia.  This band of snow was forecast to move over the
western Carolinas later in the day.

b. Sunday afternoon and evening

The area east of the mountains was becoming increasingly favorable for 
snow in the early part of the afternoon, as outlined by a Mesoscale 
Discussion issued by the Storm Prediction Center.  By 1800 UTC, the 
weak surface low was centered just west of Charleston, South Carolina, 
with a trough of low pressure extending northwest into northern Georgia
(Fig. 6).  Surface temperatures in the western Carolinas were still in 
the lower and middle 30s.  Precipitation at the surface was falling in 
the form of rain nearly everywhere in the CWA, although the rain 
changed to snow at Asheville at 1725 UTC.  The precipitation was 
diminishing in the southernmost counties where drier air circulating 
around the southern portion of the upper low was moving into the area.

The eastward progression of the deep upper-level cyclonic circulation and 
its associated cold air resulted in a lowering of the freezing level and 
downward penetration of snow in the mountains.  At the same time, the 
cold air damming pattern east of the Appalachians contributed to cooling 
temperatures near the surface.  The northeast surface winds advecting 
cold air southward from the Mid-Atlantic region were evident in the 
1800 UTC surface analysis (Fig. 6).  A northeast low-level jet greater 
than 40 kt (Fig. 10) existed at 925 mb just east of the mountains.  The 
cold air damming was not particularly strong, and the near-surface air 
moving into the region was not very cold.  Nonetheless, the temperatures 
in the lower and middle 30s were just cold enough to allow snow to reach 
the surface after the warm layer aloft was removed.  Cooling caused by 
the melting of snow also likely contributed to lowering the snow level 
and maintaining near-freezing surface temperatures.

Figure 10.  Objective analysis of 925 mb isotachs (dashed contours) and 
wind (yellow barbs) from the RUC13 at 1800 UTC on 1 March and isotachs 
(shaded contours).  A maximum wind speed of greater than 40 kt is 
located in the North Carolina Piedmont.  Click on image to enlarge.

The RUC13 initial analysis thermal profiles at 1800 UTC (Fig. 11) showed 
that significant cooling occurred in the lower troposphere since 1200 UTC.  
The AVL temperature profile was below freezing except for a very shallow 
surface layer.  The freezing level was 770 ft above the ground, so snow 
was able to reach the surface because insignificant melting occurred in 
the 33o F air.  By this time, snow was rapidly accumulating above 2900 ft 
MSL in the mountains near Asheville, and the snow level continued to drop.  
The GSP and HKY profiles displayed significant cooling since 1200 UTC, 
but CLT retained an above-freezing layer between 910 mb and 725 mb.  A 
comparison of the wind profiles in Fig. 11 with the corresponding profiles 
in Fig. 9 clearly showed the increasing depth of the surface-based cold air.  
Also, the winds above the cold layer backed to the south and southeast 
signaling the approach of the upper-level low.

Figure 11.  As in Fig. 9, except for 1800 UTC.  Click on images to enlarge.

The RUC13 objective analysis at 850 mb at 1800 UTC (Fig. 12) depicted 
a confluent northeast wind flow and nearly neutral thermal advection 
pattern over the western portion of the CWA.  Strong warm air advection 
was still occurring in the east where the weakening, but persistent, 
warm layer continued on the CLT temperature profile.  The objective 
analysis of the 295 K isentropic surface at 1800 UTC from the RUC13 
(Fig. 13) showed a well-defined region of isentropic lift from northeast 
Georgia across upstate South Carolina into central and eastern North 
Carolina.  The precipitation pattern on the mosaic of regional radars 
(Fig. 14) corresponded closely to the area of upward motion on the 
295K surface.  The precipitation was light and scattered across southern 
portions of the upstate where isentropic lift was minimal and dry air 
was circulating around the southern periphery of the upper-level low.

Figure 12.  Objective analysis of 850 mb geopotential height (green 
contours), temperature (solid and dashed brown contours), and wind 
(yellow barbs) from the RUC13 at 1800 UTC on 1 March.  Click on image 
to enlarge.

Figure 13.  Objective analysis of pressure (green contours), wind (barbs),
and vertical motion (color fill) on the 295 K isentropic surface from
the RUC13 at 1800 UTC on 1 March.  The cooler colors represent downward
motion and the warmer colors denote upward motion.  Click on image to 
enlarge.

Figure 14.  Composite radar reflectivity mosaic at 1800 UTC on 1 March.  
Click on image to enlarge.

At 2100 UTC the primary surface low was located near Myrtle Beach, South 
Carolina (MYR) (Fig. 6).  A trough of low pressure extended west to a 
weaker low center near Augusta, Georgia (AGS).  The low near AGS was the 
surface reflection of the cyclonic circulation aloft, which was located 
somewhere between AGS and Atlanta, Georgia (ATL), as seen on all levels
from 925 mb through 300 mb on the RUC13 initial analysis.  The steep 
slope of the low geopotential heights resulted in a deep layer of 
northeast to east winds across the western Carolinas and northeast 
Georgia.  Snow was observed at stations in northeast Georgia as the 
cold air associated with the low aloft moved eastward.

The cold air damming pattern held temperatures near or just above 
freezing east of the mountains throughout the afternoon.  Lower 1000-
500 mb thickness, related to the colder air aloft in the deep cyclonic 
circulation, gradually moved into the western Carolinas.  The snow 
level lowered sufficiently so that a transition of precipitation types 
occurred at most locations outside the mountains.  One notable exception 
was the North Carolina Foothills where the RUC13 925 mb temperature 
analysis and forecasts (not shown) indicated that a small area of 
relatively warm temperatures persisted throughout the afternoon.  The 
slightly warmer temperatures delayed the onset of accumulating snow and 
contributed to the smaller snow totals depicted just east of the 
mountains in Fig. 1.   Rain changed to snow at HKY around 1800 UTC, but 
it changed to rain again around 2000 UTC when the surface temperature 
increased to 36o F.  Rain continued at HKY until 0047 UTC on 2 March 
when snow returned.

The origin of the warmer air in the foothills was not clearly understood.  
Wind direction in the vertical veered from northeast to south from the 
surface to 500 mb and above, so an obvious downslope wind component to 
produce warming by subsidence did not exist.  An untested hypothesis 
suggested that the predominant cooling in the cold air damming occurred 
farther to the east near the axis of the low-level jet in Fig. 10.  The 
cooling progressed slowly along the foothills and did not lower surface 
temperatures close to freezing until late in the precipitation event.

The rain changed to snow in the Greenville-Spartanburg area around 
2130 UTC, and the snow continued for the duration of the event.  
Charlotte reported a mixture of rain and snow at 1924 UTC that changed 
to rain before becoming a mixture of sleet and rain at 2250 UTC.  Snow 
replaced the sleet and rain mixture at 2352 UTC, and continued until 
the precipitation ended after 0500 UTC on 2 March.

Click to view hourly observations and meteograms:

1) Deformation zone and trowal

The region of shearing and stretching frequently observed in the wind 
field to the west through north of a cyclonic circulation is one example 
of a deformation zone.  Deformation in the wind field in some circumstances 
can increase the temperature gradient which leads to vertical motion that 
attempts to restore thermal balance.  Forecasters often focus on the 
vertical motion in the vicinity of deformation zones to highlight areas 
of potentially significant precipitation.  An analysis of the 500 mb 
height and wind (speed, direction, and streamlines) fields and omega 
at 2100 UTC from the RUC13 revealed significant upward motion along and 
just to the east of the axis of confluent streamlines west and northwest 
of the circulation center (Fig. 15). 

Figure 15.  Objective analysis of 500 mb streamlines (solid lines with 
arrows), wind (yellow barbs) and omega (color fill) from the RUC13 at 
2100 UTC on 1 March.  Warmer colors represent upward vertical motion.
Click on image to enlarge.

Perhaps a more easily understood framework for visualizing deformation 
zone precipitation was available by considering motions on isentropic 
surfaces.  An axis of increasingly low pressure and the corresponding 
wind field extending from central North Carolina west to the mountains 
then south to central Georgia was noted on the analysis of the 300 K 
surface at 2100 UTC (Fig. 16).  Air parcels at this time were climbing 
the sharply sloping valley on the 300K surface.  The pressure decreased 
from approximately 680 mb near CLT to approximately 600 mb near ATL.  
The color shades on Fig. 16 highlight the significant upward motion 
that occurred north through west of the circulation center near AGS.  
The arc-shaped region of upward motion conformed closely to the 
precipitation indicated on the 2100 UTC composite radar image (Fig. 17).

Figure 16.  Objective analysis of pressure (green contours), wind (yellow
barbs) and omega (color fill) on the 300 K surface from the RUC13 at 
2100 UTC on 1 March.  Click on image to enlarge.

Figure 17.  Composite radar reflectivity mosaic at 2100 UTC on 1 March.  
Click on image to enlarge.

The axis of low pressures curving around the north side of the 
circulation on the sloping 300 K surface was a signature of a trowal 
(trough of warm air aloft).  The trowal has been described by a number 
of authors (e.g., Penner 1955; Martin 1999) as a key feature in 
understanding the structure, frontogenetical properties, and the 
associated three-dimensional air flow of maturing extratropical cyclones.

The ascending airstream in the trowal remained over the region until late 
at night as the trowal pivoted around the circulation center.  This 
occurred because the deep cyclonic circulation moved just south of the 
WFO GSP CWA then turned toward the northeast.  Precipitation continued 
until shortly after 0500 UTC on 2 March in the Greenville-Spartanburg 
area and until approximately 0900 UTC on 2 March at Charlotte.

The 2025 UTC infrared satellite image (Fig. 18) was enhanced to show key 
features of the primary conveyor belts comprising the maturing cyclone.  
[Refer to Carlson (1991); Moore et al. 2005; and Schultz (2001) for more 
detailed information regarding conveyor belts in an extratropical cyclone.]
The warm conveyor belt flowed northward from the Bahamas and ascended to 
the comma cloud head where it curved anticyclonically and entered the 
high level jet stream flow.  A branch of the warm conveyor belt did not 
rise as high as the primary branch while it curved cyclonically to wrap 
around the low center.  This air comprised much of the trowal airstream.  
A much lower airstream, the cold conveyor belt, had its origins in the 
cold air over the Mid-Atlantic States.  The cold conveyor belt flowed 
westward, north of the surface low pressure system, and then curved 
cyclonically around the low center.   The dry conveyor belt originated 
at high levels northwest of the cyclone then descended and approached 
the surface to the south of the low pressure center.  Some of the dry 
air wrapped cyclonically around the circulation while another branch 
curved anticyclonically and moved southeast as it neared the surface 
cold front.

Figure 18.  GOES-12 infrared satellite image at 2025 UTC on 1 March, 
showing the warm conveyor belts (red), cold conveyor belt (blue), and 
dry conveyor belt (brown).  Click on image to enlarge.

2) Thundersnow

One of the interesting features of this winter precipitation event was 
the occurrence of thundersnow from Upstate South Carolina northeast into 
the Piedmont of North Carolina.  Forecasters at GSP were anticipating
thundersnow and enhanced snowfall rates to move across portions of the
Piedmont of the Carolinas after seeing observations of thundersnow 
across north Georgia and reading the Mesoscale Discussion issued by
the SPC at 2040 UTC on 1 March.  The first report of thundersnow was 
made by the observer at the Charlotte – Douglas International Airport 
at 2352 UTC on 1 March.  Thundersnow was also observed at the Greenville– 
Spartanburg International Airport between 0129 UTC and 0228 UTC on 
2 March.  Cloud-to-ground lightning was detected between 0110 UTC and 
0200 UTC.  Anecdotal evidence suggested that thundersnow continued across 
the area from northeast of Greenville, South Carolina, to Shelby and 
Lincolnton, North Carolina, until at least 0300 UTC on 2 March. 

Click here to view a 1.11 mb windows media video of thundersnow.  
The video was shot by Caitlin Rubow north of Dallas (Gaston County), 
North Carolina, around 9 pm on 1 March 2009.  Used by permission.

Click here to view a 26 frame Java loop of Composite Reflectivity from 
the KGSP radar from 0045 UTC to 0230 UTC 2 March 2009.

The lightning was produced in a region of enhanced instability aloft.  
A cross section from the 0100 UTC RUC13 analysis (Fig. 19), extending 
from Kentucky southeast to a position off the South Carolina coast, 
depicted a significant zone of rising motion (omega) sloping upward 
from the South Carolina coast toward the mountains.  Just east of the 
mountains, a small upward motion maximum coincided with an elevated 
zone of instability denoted by the negative values of saturated 
geostrophic potential vorticity.  The upper air sounding taken at 
Greensboro, North Carolina (GSO), at 0000 UTC on 2 March (Fig. 20) 
showed an environment that was favorable for convectively enhanced 
precipitation.  Although the location of the sounding was 75 to 
150 miles to the northeast of the thundersnow area, it included many 
features seen in other thundersnow events (Market et al. 2006).  The 
most unstable parcel level was at 734 mb, which was approximately 30 mb 
above the top of the frontal temperature inversion.  Significant drying 
occurred above 650 mb which indicated a conditionally unstable layer.  
In fact, parcels lifted from above the top of the inversion were unstable 
and positively buoyant with Convective Available Potential Energy of 
311 J kg-1, indicating that upright convection was possible.  The most 
unstable parcel originated below the -1000 C level of the sounding, 
and its equilibrium level was above the -2000 C level, which indicated 
that the environment was capable of supporting cloud-to-ground lightning 
production as set forth by the composite thundersnow sounding of 
Market et al. (2006).  The 700 – 500 mb temperature difference was 
1900 C, which was at the upper part of the range of temperatures observed 
in other thundersnow events over the Southeastern United States 
(Hunter et al. 2001).

Figure 19.  Vertical cross-section from eastern Kentucky to a point just 
east of Myrtle Beach, depicting omega (yellow contours) and saturated 
geostrophic potential vorticity (contoured shades with negative values 
in blue) from the RUC13 at 0100 UTC on 2 March.  Click on image to 
enlarge.

Figure 20.  Skew-T, log P diagram for the upper air sounding taken at GSO 
at 0000 UTC on 2 March.  The red line is the temperature sounding and the 
green line is the dewpoint sounding.  A hodograph is shown in the upper 
right corner of the figure.  A table of convective parameters is included 
at the bottom of the figure.  Click on image to enlarge.

The instability that supported the lightning-producing convective storms 
was probably caused by the intrusion of the dry conveyor belt from the 
southwest into the moist air mass.   Several studies have identified the 
leading edge of the "dry slot" as a region of convective destabilization.  
Nicosia and Grumm (1999) showed that stability can decrease when the dry 
conveyor belt rides over moist air below.  Upward motion tapping the 
potentially unstable air likely resulted in the development of convective 
updrafts sufficiently strong to produce thunderstorms and enhanced snowfall 
rates.  The axis of relatively high snow accumulations seen in Fig. 1 more 
than likely owed its existence to the convectively enhanced precipitation.

Although the environment supported upright convection during the evening 
hours, individual convective towers were not discernable on the Weather 
Surveillance Radar at the Greenville-Spartanburg International Airport 
(the KGSP radar).  The precipitation was loosely organized in two bands 
stretching from southwest to northeast (Fig. 21).  One band was located 
from extreme northeast Georgia, across Oconee County, South Carolina, to 
the eastern part of the North Carolina Mountains across Transylvania, 
Henderson, and Buncombe counties.  The other band was situated across the 
southern half of Greenville County, South Carolina, and stretched across 
Spartanburg and Cherokee counties, to Cleveland and Lincoln counties in 
North Carolina.  It was the second band in which the thundersnow occurred. 
The thundersnow band contained some reflectivity greater than 40 dBZ on 
the 0.5 degree scan, generally at a level below 2,000 feet AGL.  The higher 
reflectivity probably resulted from aggregates of partially melted dendritic 
snow crystals.  Above that level, reflectivity was less than 40 dBZ and 
radar echo tops showed little variation across the band, thus it was 
difficult to detect stronger convective cells that might produce lightning.

Figure 21.  KGSP base reflectivity at (A) 0.5 degrees, (B) 1.3 degrees, 
and (C) 2.4 degrees, and (D) echo top, for the 0101 UTC volume scan on 
2 March.  The color table on the left is for the reflectivity images and 
the color table on the right is for the echo top image.  The radar
images were created using GRLevel2 Analyst.  Click on image to enlarge.

c. Sunday night and Monday morning

The two low pressure systems seen in Fig. 6 moved northeast away from 
the Carolinas late Sunday night and early Monday morning.  By 1200 UTC 
2 March, a single surface low pressure was centered off the New Jersey 
coast.  The 500 mb cutoff low had evolved into an open, sharp trough that 
was moving offshore.  In the wake of the storm, the sky cleared across 
the western Carolinas and extreme northeast Georgia allowing the Aqua 
MODIS satellite to capture the image in Fig. 21.  The image was obtained 
during an early afternoon pass, so warming temperatures had already 
eliminated the snow in locations where only minor accumulations occurred.  
Close examination of the figure revealed the lack of significant snow 
in the far western portion of the CWA and in the North Carolina Foothills.  
Lesser snow accumulations were also seen in the French Broad River Valley 
in the vicinity of Asheville.  Smaller snow accumulations in the valley 
were probably caused by above freezing surface temperatures in the lower 
elevations that delayed the downward penetration of snow and, perhaps, by 
a precipitation shadow affect. 

Figure 22.  Aqua MODIS image from approximately 1830 UTC 2 March 2009.  
The early afternoon sun caused melting where only small amounts of snow 
accumulated the previous afternoon and evening.  Nearly snow-free areas 
are in the foothills and in the lower elevations of the far western 
counties.  The French Broad River valley in the vicinity of Asheville 
is near the center of the image.  Image:  Space Science and Engineering 
Center, UW-Madison.  Image overlays:  National Environmental Modeling 
and Analysis Center, UNC Asheville.  Click on image to enlarge.

4.  Summary 

A low pressure system traveling across the southeastern United States 
from 28 February until 2 March 2009 produced rain and snow across 
portions of the western Carolinas and extreme northeast Georgia.  Snow 
accumulations ranged from a trace in far western counties to 16 inches 
at Mt. Mitchell, North Carolina.  Some Piedmont locations in North and 
South Carolina received 6 to 10 inches of snow.  Two sources of cold 
air contributed to the transition from rain to snow while the low 
pressure system traversed the area:  1) Cold air accompanying the deep 
upper-level low, and 2) Near-surface cold air supplied by the high 
pressure system near the Great Lakes.  Vertical motion related to highly 
ageostrophic flow aloft contributed to the significant and widespread 
precipitation.  The duration of precipitation was enhanced by the slow 
eastward progression of the trowal airstream as it pivoted around the 
upper-level low.  Warmer air along the Foothills delayed the transition 
to snow thus reducing snow accumulations.  Only small snowfall totals 
were observed in some of the far western counties where the westward 
penetration of the cold air damming air mass was blocked by higher 
terrain.  Snowfall rates in the Piedmont were enhanced Sunday evening 
by an unstable atmosphere that developed as the deep upper-level low 
moved just south of the area.

Photo Gallery

The view from Poga Mountain, near the Flat Springs community in 
northern Avery County, North Carolina, on the morning of 2 March 2009.  
Image by Dr. Baker Perry.

Light snow falls near the Clemson Marina around 5:15 pm on Sunday, 
1 March 2009.  Images by Ted Paskiewicz.

This was the view outside the National Weather Service Office at the
Greenville-Spartanburg Airport on the morning of 2 March 2009.  Officially,
4.4 inches fell at the GSP Airport during the event.

More views of the new-fallen snow around Lincolnton, North Carolina,
just after sunrise on 2 March 2009.

New snow on the morning of Monday, March 2, in the Five Forks area of 
Simpsonville, South Carolina.

References

Bennetts, D. A., and B. J. Hoskins, 1979: Conditional symmetric instability —
     A possible explanation of frontal rainbands. Quart. J. Roy. Meteor. Soc., 
     105, 945–962.

Carlson, T. N., 1991:  Mid-Latitude Weather Systems.  Harper-Collins Academic, 
     London, UK, 507 pp.

Hunter, S. M., S. J. Underwood, R. L. Holle, and T. L. Mote, 2001:  Winter 
     lightning and heavy frozen precipitation in the southeast United States.  
     Wea. Forecasting, 16, 478-490.

Market, P. S., A. M. Oravetz, D. Gaede, E. Bookbinder, A. R. Lupo, C. J. 
     Melick, L. L. Smith, R. Thomas, R. Redburn, B. P. Pettegrew, and A. E. 
     Becker, 2006:  Proximity soundings of thundersnow in the central United 
     States.  J. Geophys. Res., 111, D19208. 

Martin, J. E., 1999: Quasigeostrophic forcing of ascent in the occluded sector 
     of cyclones and the trowal airstream.  Mon. Wea. Rev., 127, 70–88.

McCann, D.W., 1995: Three-Dimensional Computations of Equivalent Potential 
     Vorticity.  Wea. Forecasting, 10, 798–802.

Moore, J. T., 1993:  Isentropic Analysis and Interpretation.  National Weather 
     Service Training Center, Kansas City, MO, 99 pp.

Moore, J. T., Ng, S., and C. E. Graves, 2005:  The role of conveyor belts in 
     organizing processes associated with heavy banded snowfall.  Preprints, 
     21st Conference on Weather Analysis and Forecasting, Washington, DC, 
     Amer. Meteor. Soc.

Nicosia, D. J., and R. H. Grumm, 1999: Mesoscale band formation in three major 
     northeastern United States snowstorms.  Wea. Forecasting, 14, 346–368.

Penner, C. M., 1955:  A three-front model for synoptic analyses.  Quart. J. Roy. 
     Meteor. Soc., 81, 89-91.

Schultz, D.M., 2001: Reexamining the cold conveyor belt.  Mon. Wea. Rev., 129, 
     2205–2225.

Acknowledgements

Blair Holloway prepared the snowfall map using ArcView GIS.  The surface 
fronts and pressure analyses were obtained from the Hydrometeorological 
Prediction Center surface analysis archive.  The thermal profiles in 
Figs. 9 and 11 were created using the Rawinsonde Observation Program 
version 5.8 for Windows. The upper air analyses and sounding plot in 
Fig. 20 were obtained from the Storm Prediction Center.  The surface 
observations plot, regional radar reflectivity mosaics, and satellite 
imagery were obtained from the RAP - Real Time Weather Data page 
maintained by the University Corporation for Atmospheric Research.  The 
radar imagery in Fig. 21 was created using the GRLevel2 Analyst software 
package.  Hourly observations were obtained from the National Climatic
Data Center.  The meteogram plots were obtained from the Plymouth State 
College Weather Center.  The Aqua MODIS image was obtained from the 
Space Science and Engineering Center at the University of Wisconsin at
Madison, with overlays from the National Environmental Modeling and
Analysis Center at the University of North Carolina at Asheville.

Special thanks are given to Eric Thomas (WBTV Charlotte), Michael Dross,
and Justin Rubow for providing the unusual thundersnow video clip.