The Difficulties Forecasting Northwest Flow Snow
in Western North Carolina


“Let it Snow, Let it Snow, Let it Snow” is more than a popular holiday song first recorded in 1945.(1) It’s a cry frequently echoed today from ski resorts, hotel owners and tourism agencies in Western North Carolina. Much of the region’s wintertime economy centers on snowfall and the tourists it brings to the area. No, not every visitor hits the slopes, but thousands of people visit The Biltmore Estate hoping to see the nation’s largest privately owned home blanketed by snowfall. Still, predicting that snowfall and accurately estimating accumulations for a given area are among the most challenging tasks in the Carolinas.

It has been my experience as a forecaster the past eight years at WLOS-TV in Asheville, NC, there are three primary ways snow can impact Western North Carolina. The first, and often the most significant in terms of accumulation, is a low pressure system filled with moisture advecting from the Gulf of Mexico into cooler latitudes. The second involves a fast-moving “clipper-system” racing from the Plains into the mountains, leaving only light to moderate snowfall. The third -- and most difficult to forecast -- is Northwest Flow Snow.

Northwest Flow Snow (NFS) for purposes of this assignment is, according to the National Weather Service office in Greenville-Spartanburg, SC, “snowfall in Western North Carolina occurring during periods of low-level upslope (northwest) flow across the southern Appalachians.” (2) Two regarded studies indicate NFS accounts for up to 25 percent of annual snowfall in the Appalachians.(3) At least 75 NFS events have been documented in Western North Carolina between 1980 and 1990, with the average number of annual events in this region dropping to about five since 1991.(4,5) Snow totals in a single event can range from a trace to feet, with the heaviest snow almost always occurring on the western-facing slopes.

According to Vince Dicarlo, meteorologist in charge of the National Weather Service office in Greenville-Spartanburg, SC, “The classical northwest flow snow is not associated with the comma head portion of an extratropical cyclone precipitation shield. Instead, the snow is associated with the lifting of the broad area of stratus and stratocumulus clouds that accompany the cold air intrusion following the passage of a low pressure system.” (6) Researchers agree the primary cause of the upward vertical motion necessary to produce northwest flow snow is the rapid increase in elevation that occurs along the Tennessee-North Carolina border.(7) In fact, elevation increases nearly 4,000 feet over 70 miles in Western North Carolina. One example is the stretch of Interstate 26 from Polk County along the South Carolina border (1,100 feet) to Madison County along the Tennessee border (4,900 feet).

Baker Perry is among few people to research NFS in Western North Carolina. He is a faculty member in the Department of Geography and Planning at Appalachian State University in Boone, NC. He concludes that NFS “typically occurs in conjunction with a surface cyclone to the east-north east and a 500 mb trough to the northeast.” (It’s been my experience that higher surface pressures are typically found west of Western North Carolina, such as in Nashville, TN, the Cumberland Plateau and even Knoxville during a NFS event.) Perry adds: “The southern Appalachians lie within the pressure gradient that results, producing low-level northwest winds that are perpendicular to the southwest-to- northeast-oriented topography. Low and mid-level moisture can be wrapped around from an intensifying surface cyclone to the northeast with this synoptic pattern.” (8) Basically, the quick change in topography provides orographic lift to extract moisture from the clouds.

Trying to predict the amount of NFS can be more than difficult. For example, during a predicted NFS event on December 18-20, 2003, the town of Banner Elk, 65 miles northeast of Asheville, picked up 25 inches (west slopes). Asheville, however, received only 3 inches, but the community of Celo some 30 miles northeast of Asheville received 30 inches (west slopes), despite its elevation of 2,700 feet compared to Banner Elk’s 3,700 feet. (9,10)

The winter of 2007 has also brought variations, including two NFS snow predictions that left Asheville with only flurries -- and forecasters with egg on their faces. Aside from personal embarrassment, consequences this year alone in the Asheville area have included schools unnecessarily letting out early or canceling altogether and parents’ workplaces disrupted by non-snow events. Because of this, there is concern some viewers will not take future snow predictions seriously. “The difficult part of the forecast is to know with confidence how much snow will fall and where significant accumulations will occur.” Dicarlo says.(11) “The terrain variability and the cellular or linear nature of the precipitation elements impinging upon the mountains create a situation in which mesoscale details cannot be known.”

The National Weather Service, however, began in the past few years classifying Western North Carolina NFS events into three categories -- post-frontal, cut-off and wrap-around.(12) The terminology can sometimes be found in forecast discussions to help a forecaster better gauge snowfall predictions. Generally speaking, it’s been my experience that all three can cause significant snow but for very different reasons. Often, the depth and size of the low associated with the post-frontal are significant, hence allowing a broad area of clouds to lift and precipitate over a larger geographic area. The cut-off low is not usually tightly wound and often occurs in late winter or early spring, dumping snow only in the highest elevations due to overall warmer temperature profiles. Nevertheless, the slow-moving tendencies can result in significant snowfall, especially at elevations above 4,000 feet. The wrap-around, basically the tail of a storm, is usually the fastest-moving of the three. That, in itself, can limit precipitation. However, the wrap-around can cause major snowfall if moisture from the Great Lakes (notably Lake Michigan) is transported, lifted and precipitated.(13)

So, how can a forecaster provide more accurate predictions (of occurrence and accumulations) regarding NFS? There are several ways, including paying close attention to 500, 700, 850 and 1000 mb charts. During a potential NFS event, it is important to know how much moisture is available, whether the cloud fields are extensive and if in- cloud temperatures are favorable for snow formation. Infrared satellite image can help with cloud height and temperature. It is also important in the Asheville area to examine the cloud fields in East Tennessee to determine thickness that will be heading toward Western North Carolina. In addition, after viewing surface temperatures at locations in Tennessee, Doppler radar could prove useful. If light, linear-type precipitation shows up, there is every reason to believe it will lift and intensify over Western North Carolina, assuming adequate temperature profiles.

Nevertheless, there is no guarantee any NFS forecast will verify well. History must be considered. Climate data show the windward sides of mountains stand to be impacted far more than leeward sides, which often experience no significant accumulation. Even the windward sides report great variations, with lower elevations sometimes receiving greater totals (as previously referenced with Celo and Banner Elk). Consequently, it is difficult to predict with any degree of certainty how much snow will fall in a particular area. Therefore, providing viewers with a range works best. For example: “Expect two to four inches of snow tonight for high mountain counties in Western North Carolina bordering Tennessee. Those of you living along the western-facing slopes could see even more snow in isolated areas.” (14) Though more vague than I’d like, it does make clear the viewer should be prepared for significant snowfall.

For forecasts to improve, technology must also improve. GFS models, especially NGM in this region, provide a reasonable warning for NFS. But, meteorologists at the Greenville-Spartanburg office of the National Weather Service agree “improved model physics and improved model resolution will result in better mesoscale and numerical guidance” to better pinpoint snow locations and accumulations. Larry Lee, the science operations officer at the National Weather Service office in Greenville-Spartanburg, says improved mapping of the high-mountain terrain is needed “to better understand the processes happening at different elevations.” He calls the current mapping “crude,” but expects improvements by 2009. (15)

Fortunately, too, more research than ever is under way. Perry continues his work, expanding NFS studies into the entire Appalachian region. Blair Holloway at North Carolina State University, meanwhile, probes links between enhanced NFS and interaction with the Great Lakes. The findings, coupled with better mapping, will result in improved forecasts. Until then, forecasters must analyze every piece of data available. It is also important that a forecaster admit on-air when a forecast busts and explain why. Equally, a forecaster should be careful declaring defeat. Dealing with NFS, I have found no accumulation on one mountaintop at 5,000 feet, but on the same day, 3 inches at lower elevations. Having a reliable weather spotter network helps a forecaster readily verify or discredit prognostications. Having knowledge, tenacity and patience will help the forecaster through the next NFS storm. So, go ahead, “Let it Snow, Let it Snow, Let it Snow."


1, 16 - “Let it Snow, Let it Snow, Let it Snow”

“Let it Snow,”

2 - Dicarlo, Vince. “Northwest Flow Snow in Western North Carolina.” National Weather Service presentation at WLOS-TV, 20 January 2006. p. 2.

3 - Schmidlin, T.W. 1992. “Does lake-effect snow extend to the mountains of West Virginia?” Proceedings of the 49th Eastern Snow Conference: 145-148.

Perry, Baker (with Konrad, Charles). 2004. “Northwest Flow Snowfall in the Southern Appalachians; Spatial and Synoptic Patterns” 61st Eastern Snow Conference: 183.

4 - Perry p.181.

5 - Holloway, Blair. 2005. “Northwest Flow Snow Events in the Southern Appalachian Mountains” CSTAR Workshop. (Slide: Notable NW Flow Events)

6, 12 - Dicarlo p. 7

7 - Dicarlo p. 3

8 - Perry p. 184

9 - Holloway p. 4, Selected Examples.

10 - “Banner Elk Community Profile,”
“Celo Community Profile,”

11 - Dicarlo p. 6

13 - Holloway. Slide: Method

14 - Siltzer, Jay. “Accuweather Forecast.” WLOS-TV. 29 January 2007. News 13 at Noon.

15 - Lee, Larry. Telephone Interview. 5 March 2007.