WRAL WeatherCenter Blog

Long-term look: A 'snowless' winter in central NC?

Posted February 10, 2015
Updated February 12, 2015


While we've had a few brushes with wintry weather this season (including some pockets of wet snow and sleet in northern parts of our area as I write this on a Tuesday morning), the question has been raised to us about how often we end up without any measurable snow during a given month or for the winter season as a whole.

I did some delving into historical statistics for the Raleigh area, and thought I'd pass along a few of the numbers that turned up. I actually looked at two different cuts from the long-term data to see how the numbers compare – if you look at all the observations available back to 1887 or if you restrict the period to the current 30-year "normals" time frame that runs from 1981-2010.

I'm posting three images here that are based on the 1981-2010 time frame, since that time period accounts a little better for patterns we've experienced most recently, and yet it is long enough to offer some reasonable stability and statistical confidence.

First, a simple look at the "normal" snowfall for that time period (meteorologists and climatologists define "normal" as more or less an average over the 30 years ending with the most recent "zero" year).

Normal Raleigh snow 1981-2010The first graph shows that our normal yearly snowfall around Raleigh is just under six inches, with most of that falling in January and February, but some notable contributions from December and March (there are also non-zero, but quite small, normals from November and April that contribute to the 5.9 inch total).

For those of you who follow statistics, you might wonder if snowfall follows a distribution that makes values like average, median, mode and so on run pretty close to one another. Well, the answer is that, on a seasonal basis, is not real far off. But taken a month at a time, the average values are skewed pretty high by the infrequent occurrence of large snowfalls, while there are lots of years with much lower totals.

This is illustrated by the median, shown in the second graph.

Median Raleigh snow 1981-2010

Median is the amount of snow such that half the years in the record had more snow than that, and half had less. You'll see that the median amount of snow on a monthly basis is much smaller than the average, because of the sizable number of times we have rather meager amounts compared to the years that we have a lot of snow. In January, the normal snowfall is 2.9 inches (influenced by years like 2000, when over 25 inches fell) while about half the time we get less than three-quarters of an inch for that month. Adding up the snow for the whole season helps to smooth all that out some, so that the median for the season is 4.3 inches, compared to the 5.9 inch normal. Also note here that the "0.0 inch" values shown for December and March actually represent trace amounts (some snow observed, but less than 0.1 inch accumulation).

So how often do we have a winter month, or a whole snow season, without any measurable snow (we're talking a trace or under one-tenth of an inch)? That pretty much counts as essentially none for most of you, so the third graph incudes months or years in which we did have some snow, but only a trace.

Percent years trace or less snow for Raleigh

As you see, for the 1981-2010 period, December and March end up with no measurable snow almost three-quarters of the time, while that happens about 40 percent of the time for January and half the time for February. So, the fact we didn't get measurable snow in December or January in Raleigh isn't very unusual historically. However, it is somewhat tougher to string together the whole November through April time frame without getting some measurable snow, as seen by the seasonal percentage of only failing to see that 13 percent of the time (or about 1 in every 7-8 years). The last time that happened, by the way, was the winter of 2005-06.

Most of these numbers are a little lower for the 1981-2010 period than the same stats covering the entire period of record back to 1887. For example, the average seasonal snowfall for that whole period is 7.4 inches, compared to the more recent 5.9. Likewise, while more recently we have a 13 percent chance of going through the season without measurable snow, the longer-term average is 9 percent, or around once in 11 years.

Part of that decrease appears to be a bit of a "squeeze" on the season, as the long-term average snow for December and March was 1.1 and 1.2 inches, respectively, while for the "normals" during that period fell to .6 and .5 inches. The most snow on record for a season was in 1892-93, when Raleigh got 31.6 inches, and there have been 12 winter seasons since 1887 with no measurable snow.

After all that, you might wonder what it means for this year. That's the problem with long-term statistics – they aren't real helpful for short-term predictions!

The fact that the percentage of entire winters that go "snowless" is much lower than any of the core snow months tells us that going through a couple of months without snow doesn't make the odds real high that we'll fail to see some in the remainder of the winter.

There are lots of historic examples of sizable snows later in the season as well. We've had measurable snow as late as April 18 (in 1983), and there was a 10-inch snowstorm on April 3 in 1915.

It appears we'll have some very cold air arrive toward Valentine's weekend, and we'll likely stay below normal for temperatures through the following week.

That doesn't offer any guarantees that a system will combine with some of that cold air in just the right way, and so far it appears mainly dry through the weekend, but it's something to watch!


Please with your WRAL.com account to comment on this story. You also will need a Facebook account to comment.

Oldest First
View all
  • James Davis Feb 26, 2015
    user avatar

    So how is that 'snowless' winter working out in NC?

  • Andy Barnhart Feb 12, 2015
    user avatar

    Seems overly complicated. What is the median date of the last snowfall? If we are past it or it is within the forecast window and there is no likely snow forecast for the entire extended forecast, then it is likely we will have a snow less winter based on past history.

  • Bill Price Feb 10, 2015
    user avatar

    No he doesn't. Outliers should not be removed unless there is concrete evidence of said outliers not be representative of the sample (e.g. equipment malfunction).

  • Paul Donovan Feb 10, 2015
    user avatar

    You need to remove the outliers (eg 25 inches in Jan 2000) to get any meaningful statistical analysis