Convective Snow Showers Across the Upstate
and Western Piedmont on 16 February 2013
Patrick D. Moore
Newly fallen snow sticks to the trees and on the ground near Wellford, South Carolina, during the afternoon of Saturday, 16 February 2013. Image courtesy of Mike Morgan and used by permission.
Author's Note: The following report has not been subjected to the scientific peer review.
Convective snow showers developed over the upper French Broad River valley and Balsam Mountains of North Carolina during the early part of the afternoon on Saturday, 16 February 2013. The snow showers spread east across Upstate South Carolina and the southern foothills and Piedmont of North Carolina during the late afternoon and early evening as they coalesced around a mesoscale area of low pressure to the lee of the Appalachians. Accumulations of one to three inches were common across the area from Transylvania and Henderson counties in North Carolina, the northern part of the Upstate, and most of the Charlotte metropolitan area. The convective snow bands produced the greatest accumulation across the area from Gastonia, North Carolina, and the west side of Charlotte, North Carolina, to Rock Hill, Union, and Chester in South Carolina (Fig. 1). The most snow fell at some high elevations of the North Carolina mountains, aided by a favorable northwest flow of moisture that began early in the day. Several observers also reported "thundersnow" during the period of most intense precipitation. Thundersnow was also observed at the Greenville-Spartanburg International Airport (GSP), the Charlotte Douglas International Airport (CLT), and the Rock Hill - York County (South Carolina) Airport (UZA) (Table 1).
Fig. 1. Snowfall accumulation across the western Carolinas for the 36 hour period ending at 1200 UTC on 17 February 2013. Click on image to enlarge.
Table 1. Surface observations taken at selected sites across the western Carolinas during the calendar day of 16 February 2013. Click on the four-letter identifier to view the observations.
Click below to view weather observations on 16 February 2013.
The events of 16-17 February 2013 were brought about by a conjunction of a shallow, cold, conditionally unstable air mass and the arrival of a short wave trough at 500 mb during the time of peak heating. In many respects, the event resembled one of ordinary showers and thunderstorms developing on a summer afternoon over the higher terrain of western North Carolina and then spreading east across the foothills and Piedmont late in the day, except in this case it was cold enough for the precipitation to reach the ground as snow.
2. Synoptic Features
A cold front moved across the western Carolinas during the evening of 15 February and ushered in an arctic air mass that settled over the region on 16 February. No precipitation accompanied the passage of the front. However, light precipitation developed on the cool side of the boundary across northeast Georgia, Upstate South Carolina, and the western Piedmont of North Carolina, mainly along and east of Interstate 85, between 0600 UTC and 0900 UTC on 16 February. Most of the initial precipitation fell as rain, but then mixed with or changed to snow after sunrise as the boundary layer continued to cool. Light precipitation and cloudiness were prevalent across the area to the east of the Blue Ridge during most of the morning hours, which prevented temperatures from warming appreciably.
A highly amplified upper pattern was present at 500 mb across most of North America at 1200 UTC on 16 February (Fig. 2). The axis of an upper trough was oriented roughly along a line from Lake Michigan to the lower half of the Mississippi River. Water vapor imagery showed the characteristic swirl of a cyclonic vorticity maximum moving around the bottom of the trough from Arkansas into northern Mississippi and western Tennessee. The 850 mb analysis showed evidence of a cold front aloft extending from Delaware, across the central Carolinas, to southern Alabama (Fig. 3). Winds were northwest behind the boundary from the western Carolinas to the upper Midwest, with an orientation nearly perpendicular to the isotherms indicating cold advection. Temperatures at 850 mb ranged from -2°C across the Piedmont to -8°C along the Tennessee border. At the surface, the cold front had moved off the East Coast (Fig. 4).
Fig. 2. Storm Prediction Center (SPC) objective analysis of 500 mb geopotential height (dm; dark gray contours), temperature (oC; dashed red contours), and wind (kt; barbs) at 1200 UTC on 16 February 2013. Click on image to enlarge.
Fig. 3. SPC objective analysis of 850 mb geopotential height (dm; dark gray contours), temperature (oC; dashed contours, red above and blue below 0 oC), dewpoint (greater than 8 oC; green contours), and wind (kt; barbs) at 1200 UTC on 16 February 2013. Click on image to enlarge.
Fig. 4. Hydrometeorological Prediction Center (HPC) regional surface analysis of sea level pressure (mb; black contours) and fronts (traditional symbols) at 1200 UTC on 16 February 2013. Observations are indicated by the traditional station model plot. Click to enlarge.
By late morning, the approach of the upper trough helped to induce weak cyclogenesis over western South Carolina, as seen on the surface analysis at 1500 UTC (Fig. 5). The weak low pressure center persisted to the lee of the Appalachians through the late afternoon and helped to increase surface moisture convergence east of the mountains. The veil of thicker mid- and high-level cloudiness moved off to the east between 1500 UTC and 1800 UTC, allowing some limited surface heating over the North Carolina mountains and the western part of Upstate South Carolina.
Fig. 5. As in Fig. 4, except for 1500 UTC on 16 February 2013. Click to enlarge.
By 1800 UTC, the atmosphere was synoptically and thermodynamically favorable for shallow moist convection. The cyclonic vorticity maximum had moved across the Southeast to a position over north and central Georgia (Fig. 6). Differential positive vorticity advection, usually associated with upward vertical motion in the atmosphere, was maximized over the northern part of Upstate South Carolina and the central mountains of North Carolina. Frontogenesis at 850 mb and differential divergence between 850 mb and 250 mb were also present in the same relative area (Fig. 7). Although shallow, the boundary layer air mass was unstable with a lapse rate in the 0-3 km layer of approximately 8 oC km-1 and a most-unstable convective available potential energy (CAPE) of approximately 100 J kg-1, indicating that shallow convection was possible. An initial hour profile of temperature and dew point from the Rapid Refresh (RAP) model at 1800 UTC (Fig. 8) at GSP showed the steep temperature lapse rate that existed from the surface up to 600 mb (about 14000 feet above ground level [AGL]) and a freezing level around 1500 feet AGL. The height of the -20 0C level was approximately 11000 ft AGL.
Fig. 6. SPC objective analysis of 500 mb geopotential height (dm; black contours), vorticity (105 s-1; dashed black contours and color fill), with 700-400 mb differential positive vorticity advection (109 s-1; blue contours) at 1800 UTC on 16 February 2013. Click on image to enlarge.
Fig. 7. SPC radar reflectivity mosaic with objective mesoscale analysis of (a) 850 mb geopotential height (m; black contours), temperature (oC; dashed blue contours), wind (kt; barbs), and Pettersen Frontogenesis (oK 10-2 km-1 3 hr; purple contours), (b) 850 mb convergence (10-5 s-1; red contours), 250 mb divergence (10-5 s-1; purple contours), and 850-250 mb differential divergence (dashed black contours and color fill), (c) 0-3 km lapse rate (oC km-1; contours, red greater than 8 and color fill), and (d) most unstable CAPE (J kg-1; orange contours) and lifted parcel level (m AGL; black contours and color fill) at 1800 UTC on 16 February 2013. Click on each image to enlarge.
Fig. 8. RAP initial hour model profile of temperature, dewpoint, and wind at GSP at 1800 UTC on 16 February 2013. The dashed green line represents the hypothetical path of a parcel lifted from the surface. Click to enlarge.
3. Radar Observations
The National Weather Service Doppler radar at the GSP airport (the KGSP radar) detected convective initiation across northern Transylvania, northwest Henderson, and southwest Buncombe counties in North Carolina on the east- and southeast- facing slopes of the Balsam Mountains around 1500 UTC (Fig. 9), perhaps in response to favorable heating after some brief clearing seen in the visible satellite imagery. Additional convection began on the southeast-facing slopes of the South Mountains and the Blue Ridge of South Carolina by 1600 UTC, then expanded across northern Greenville County, South Carolina, through 1700 UTC. Radar echo tops were generally under 15000 feet AGL until 1700 UTC, around which time the back edge of the high cloudiness moved across the region where the convection had developed. The removal of the higher clouds resulted in more surface insolation which increased the buoyancy of surface air parcels, which may have allowed additional convective initiation along and east of the Blue Ridge. By 1800 UTC, numerous showers were noted from Pickens County, South Carolina, and the northern part of Greenville County, northward across the foothills of North Carolina.
Fig. 9. Composite reflectivity from the KGSP radar at (a) 1456 UTC, (b) 1604 UTC, (c) 1702 UTC, and (d) 1803 UTC on 16 February 2013. Note the change from clear air mode to precipitation mode between 1702 UTC and 1803 UTC. Click on image to enlarge.
Some of the showers on the southern flank of the convection grew to an appreciable size around 1800 UTC. At 1754 UTC, a cell over Pickens County had a core of reflectivity greater than 40 dBZ that extended to nearly 12000 feet above ground level (AGL) seen on the 5.1 degree elevation scan from the KGSP radar (Fig. 10). Compared to the RAP temperature profile in Fig. 8, the cell was tall enough to potentially have enough charge separation for cloud-to-ground lightning. The dual-polarized beam from the KGSP radar allowed for a determination of the most likely precipitation type in the showers occurring over Pickens County (Fig. 11). The KGSP radar showed a low differential reflectivity (ZDR, less than -0.1 dB), a low specific differential phase (KDP, less than 0.15 deg km-1), and a very high correlation coefficient (CC, greater than 0.99), which suggested the precipitation was falling as snow at the height of the radar beam, which was approximately 1600 ft to 2200 ft AGL. The precipitation over northern Greenville County was probably falling as a combination of rain and snow, suggested by the relatively lower reflectivity and the less uniform appearance of the CC product.
Fig. 10. KGSP base reflectivity on the (a) 0.5 degree, (b) 1.3 degree, (c) 3.1 degree, and (d) 5.1 degree scans at 1754 UTC on 16 February 2013. The height of the beam as it passes through the small area of 40 dBZ reflectivity to the northeast of point "KLQK" was approximately 12000 ft AGL on the 5.1 degree scan in (d). Click on image to enlarge.
Fig. 11. KGSP 0.5 degree scan of (a) base reflectivity (dBZ), (b) ZDR (dB), (c) KDP (deg km-1), and (d) CC at 1754 UTC on 16 February 2013. Click on image to enlarge.
The first cloud-to-ground lightning strike was detected northwest of the KGSP radar at approximately 1834 UTC. At that time, the KGSP radar showed a west- to-east elongated band of showers moving toward the radar (Fig. 12). The outflow boundary reached the airport at 1838 UTC, shifting the wind direction to northwest and lowering the dew point temperature. Ten minutes later, the wind had shifted to north at 14 kt. Precipitation began at GSP in the form of snow at 1851 UTC with visibility quickly dropping to one mile, then heavy snow fell from 1859 UTC to 1951 UTC at the rate of about two inches per hour. By 1930 UTC, the area of convective snow resembled a mesoscale convective system (MCS) with a well-defined outflow boundary and cold pool moving toward the southeast into northern Laurens and southern Spartanburg counties in South Carolina (Fig. 13). Meanwhile, the precipitation moving southeast from the Blue Ridge in North Carolina consolidated into an elongated region with a stratiform appearance stretching from the convective elements near the Greenville-Laurens-Spartanburg county lines north-northeast across the western Piedmont of North Carolina.
Fig. 12. KGSP 0.5 degree scan of (a) base reflectivity (dBZ) and (b) base velocity (kt) at 1834 UTC on 16 February 2013. One cloud-to-ground lightning flash is indicated by the "-" sign below the "e" and "n" in "Greenville." Click on image to enlarge.
Fig. 13. As in Fig. 12, except for 1930 UTC on 16 February 2013. Click on image to enlarge.
Through 2000 UTC, the broad arc of convective cells on the leading edge of the cold pool continued to move southeast across northwestern Laurens County and southeastern Spartanburg County, while the parallel stratiform region extended north and northeast to the vicinity of Lake Norman in the western Piedmont of North Carolina (Fig. 14). The stratiform region was broken only by two brief cells of higher reflectivity (i.e. greater than 35 dBZ) over southeastern Cleveland County and western Gaston County, North Carolina. Three more cloud-to-ground lightning flashes were detected near Cross Keys in southeastern Union County, South Carolina, around 2045 UTC. Visible satellite imagery at 2045 UTC showed evidence of southwest to northeast flow through the convection over Upstate South Carolina by the wavy appearance of the clouds along the border with North Carolina (Fig. 15). To the west, horizontal convective rolls and closed cell convection were noted over much of the Tennessee Valley and Cumberland Plateau, while to the east the back edge of the thicker cloudiness associated with the baroclinic zone over the eastern half of North Carolina was seen just east of the Charlotte metro area. A radar mosaic at 2059 UTC showed numerous showers over eastern Kentucky and middle Tennessee associated with the closed cell convection across the Cumberland Plateau, moving in the direction of the North Carolina mountains (Fig. 16). A RAP initial hour profile of temperature, dew point, and wind over northern South Carolina at 2100 UTC (Fig. 17) also showed the southwest to northeast flow in the layer between 500 mb and 600 mb, which roughly corresponded to the tops of the convective cells.
Fig. 14. KGSP composite reflectivity from the 2000 UTC volume scan on 16 February 2013. The Lake Norman area of North Carolina is located on the border between Catawba County and Iredell County, below and right of the last 'a' in North Carolina. Click on image to enlarge.
Fig. 15. GOES-13 visible satellite imagery at 2045 UTC on 16 February 2013. The convective part of the MCS was located above and to the right of "A", while "B" denotes the parallel stratiform region with the wave clouds shown by a washboard appearance. Horizontal convective rolls are shown near "C" and closed cell convection near "D". A large area of precipitation is shown across the eastern half on North Carolina near "E". Click on image to enlarge.
Fig. 16. Regional radar reflectivity mosaic at 2059 UTC on 16 February 2013. Locations "A" through "E" are identified in Figure 15. Click on image to enlarge.
Fig. 17. As in Figure 8, except for UZA at 2100 UTC on 16 February 2013. Click to enlarge.
The snow reached Gastonia, North Carolina, at 2030 UTC and then Charlotte, North Carolina, around 2115 UTC. Heavy snow began at the Charlotte-Douglas International Airport at 2130 UTC and was accompanied by thunder from 2143 UTC through the remainder of the hour. The precipitation began as light rain with thunder at Rock Hill, South Carolina, at 2135 UTC and quickly changed to moderate snow with thunder at 2143 UTC. The observations of thundersnow at Charlotte and Rock Hill were associated with another flurry of lightning activity to the west and southwest of Charlotte, which peaked around 2146 UTC with four cloud-to-ground flashes in a five minute period, all within the parallel stratiform region of the MCS (Fig. 18). It is speculated that the charge separation required for the cloud-to-ground lightning flashes occurred in the convective cells, located by this time over western Fairfield County, South Carolina, with the charged hydrometeors advected toward the northeast by the flow through the upper part of the MCS.
Fig. 18. KGSP base reflectivity on the (a) 0.5 degree, (b) 0.9 degree, (c) 1.3 degree, and (d) 1.8 degree scans at 2146 UTC on 16 February 2013. The locations of the four cloud-to-ground lightning flashes are shown by the "-" symbols. Click on image to enlarge.
The snow remained over a corridor parallel to Interstate 77 from western Mecklenburg County, North Carolina, to eastern York County, South Carolina, and northeastern Chester County, South Carolina, for over three hours through about 0000 UTC on 17 February. The relatively long duration was due to the movement of the weak surface low and mid level vorticity center forcing the axis of the stratiform region to pivot over the area near Rock Hill. As a result, the highest snow accumulations were observed from York and Rock Hill to Edgemont (Chester County), South Carolina. The back edge of the trailing stratiform precipitation region exited Union County, North Carolina, around 0130 UTC on 17 February. A few bands of snow showers lingered into the late evening hours over the northwest flow upslope areas of the North Carolina mountains.
Sky began to clear from the west in the wake of the short wave during the late afternoon on 16 February, allowing for some melting to take place over the western Upstate of South Carolina and the French Broad Valley of North Carolina. Sky remained clear overnight, allowing temperatures to drop well below freezing across the region and preventing the snow from melting further. On Sunday morning, 17 February, the Terra MODIS satellite passed overhead between 1551 UTC and 1601 UTC (Fig. 19). The remaining snow cover across the southern Piedmont of North Carolina and the northern Upstate of South Carolina appeared as a whitish fan-shaped region that tapered down to a corridor across the northern part of the Midlands and coastal plain of South Carolina. About four hours later, the Aqua MODIS satellite passed overhead between 1910 UTC and 1922 UTC (Fig. 20). The amount of melting was dramatic and most of the remaining snow cover was limited to areas to the west and south of Charlotte, North Carolina.
Fig. 19. Terra MODIS imagery from 1551 UTC to 1601 UTC on 17 February 2013. Click on image to enlarge.
Fig. 20. Aqua MODIS imagery from 1910 UTC to 1922 UTC on 17 February 2013. Click on image to enlarge.
Several convective snow showers developed over the upper French Broad River valley of North Carolina and moved east-southeast off the Blue Ridge in South Carolina during the afternoon of Saturday, 16 February 2013. The convection was enhanced by a favorable conjunction of low- and mid-level forcing on the synoptic scale arriving during the time of maximum heating. The showers appeared to organize into an MCS complete with a cold pool/outflow boundary interacting with a conditionally unstable boundary layer, and a parallel stratiform precipitation region. The convection was deep enough and updrafts sufficient to produce several cloud-to-ground lightning flashes, mostly occurring with observations of moderate to heavy snow. Precipitation began in many places as a brief period (less than ten minutes) of light rain or a rain/sleet/snow mix, then changed to snow as the intensity of the precipitation increased to moderate and/or heavy. Several locations experienced heavy snow falling at the rate of nearly two inches per hour.
The surface analysis graphics were obtained from the archive at the Hydrometeorological Prediction Center. The upper air analyses and mesoscale analyses were obtained from the archive at the Storm Prediction Center. Most satellite imagery and radar mosaics were obtained from the University Corporation for Atmospheric Research. Surface observations were obtained from the National Climatic Data Center. The NASA Aqua and Terra MODIS satellite imagery was obtained from the Space Science and Engineering Center at the University of Wisconsin - Madison. Jeffrey Taylor created the snow accumulation map. Mike Morgan provided the image of snow cover.
Use of lightning data by the NWS provided through a license agreement with Vaisala/GAI. Reference to any specific commercial products, process, or service by trade name, trademark, manufacturer, or otherwise, does not constitute or imply its recommendation, or favoring by the United States Government or NOAA/National Weather Service. Use of information from this event review shall not be used for advertising or product endorsement purposes.