Yale Climate Connections - Jeff Masters Weather Blog

What’s behind this winter’s U.S. snow drought? » Yale Climate Connections


It’s not that the United States has been entirely bereft of snow during what’s likely to end up as the nation’s warmest winter on record. But for most of the traditionally snowy swaths from the northern Great Plains to the Northeast, there’s been a startling lack of winter storms in 2023-24. The result: widespread bare ground in midwinter and what could end up being some all-time lows for seasonal snowfall.

On top of that, there have been a couple of bizarre cases of extremely dry or wet snows, both of which threw monkey wrenches into the forecast and left many snow lovers crestfallen.

At weather.com, Jonathan Erdman pulled together some of the noteworthy seasonal stats for a February 23 writeup. Erdman noted that the average snow extent across the contiguous U.S. on that date was a paltry 17%. That’s well short of the average to date of 37%, and the lowest for any February 23 in satellite-derived data going back to 2004.

Throughout much of this winter, U.S. snow cover has hovered at or near record-low extents in the 21-year satellite database. The biggest exception came after a mid-January pair of fast-moving storms that left a picture-perfect winter landscape from the Upper South (Nashville picked up 7.6 inches of snow, more than 150% of a typical winter’s entire snowfall) to the mid-Atlantic (Washington, D.C., recorded 7.8 inches, including the city’s first inch-plus snow on a calendar day in two years).

Meanwhile, the Upper Midwest and Great Lakes states — typically plastered with snow — are seeing major deficits for the winter to date. As of February 24, some of the biggest losers (counting only those years since consistent snow measurement began at each site) included:

Location Observed thru Feb. 25 (in.) Average thru Feb. 25  (in.) % of average Deficit (in.)               
Fargo, North Dakota 8.5
(2nd-lowest in 131 years)  
37.1 23% 28.6
Minneapolis, Minnesota 14.2
(7th-lowest in 140 years)
38.4  37%                      24.2
Duluth, Minnesota 18.5
(lowest in 76 years)
65.7 28% 47.2  
Syracuse, New York 34.5  
(lowest in 77 years)  
101.8  34% 67.3

Ice extent across the Great Lakes has also dipped to record lows for much of January and February, as discussed by Jeff Masters in a post on February 16.

El Niño turns up the burners, with an assist from global warming

El Niño is clearly a big driver behind this winter’s dearth of Snow Belt snow. A periodic warming of the eastern tropical Pacific, El Niño torques winter weather across much of the Western Hemisphere, especially during strong El Niño events such as the one we’ve had in 2023-24.

The classic playbook for El Niño winters is milder and drier than average across the northern U.S. and relatively cool and wet toward the Sun Belt. This winter, it’s easy to see how the playbook has gotten skewed in a way entirely consistent with human-caused warming. Nearly all of the contiguous U.S. has been milder than average, with the most exceptional warmth following the El Niño mold across the Midwest and Northeast and into southern Canada.

Globally, we’ve just had the warmest average surface temperatures for both December and January in records dating back to 1880, and February may follow suit. As global temperatures rise, the biggest spikes atop that trend usually occur during El Niño years.

Based on forecasts extending through the last few days of February, a number of U.S. states and cities — and perhaps the contiguous U.S. as a whole — will wrap up their warmest meteorological winter, defined as December through February, in more than 125 years of recordkeeping. We’ll have a full report when NOAA releases the final numbers on March 8.

Two extreme modes of snow that gave forecasters fits

Not only has this winter’s U.S. snowfall been sparse; at times, it’s been just plain odd. Consider what happened along Colorado’s populous Front Range (including Denver, Boulder, and Fort Collins) on February 3. As a strong jet stream dove into the central U.S., wrapping moisture toward the Rockies, it was clear that the Front Range might get a major winter storm. But there were also failure modes in sight, including mild temperatures more akin to March or April than midwinter.

For days, computer models vacillated on whether the storm’s focus might be just east of the Front Range cities or atop them, and whether the storm would transition quickly to snow or instead kick off with an extended round of cold rain — something that’s virtually unheard of along the Front Range in early February.

At the National Weather Service in Denver, forecasters bemoaned the situation, including in one technical discussion issued less than two days before the event:

…this is easily one of the more difficult forecast scenarios for the Denver metro (and northeast Colorado as a whole) in quite some time … The past few model cycles have left our forecast “clear as mud” at this point.

The storm’s complex dynamics ended up favoring the big Front Range cities after all, and there was enough rain at the outset to astound many long-time weather watchers. The storm also featured snow-to-liquid-water ratios on the order of 5:1 – meaning five inches of snow containing an inch of moisture – compared to a more typical Colorado ratio of around 15:1. (See Figure 1 below, as well as the interactive website at Saint Louis University.)

Map of the snow-to-liquid ratio climatology stratified by NWS Forecast Zone, showing that interior and northern areas of the U.S. typically have drier snow than coastal and southern areas.
Figure 1. Typical ratios of snow-to-liquid-water in various parts of the United States. The numbers on the left axis correspond to the ratios; for example, “15” means a 15:1 ratio, or 15 inches of snow for every inch of liquid. (Image credit: Saint Louis University/CIPS)

In Fort Collins, where the storm was mostly a bizarre midwinter rain with a mere one inch of sodden snow, the total precipitation of 1.66 inches on February 3 was not only the most for any February day in 131 years of recordkeeping (far exceeding 1.02 inches from Feb. 27, 1918), it was also the most for any entire February, just above the 1.65 inches in Feb. 1912. Also reported:

  • Boulder:  1.58 inches of precipitation and 6.9 inches of snow. Wettest February day on record (topping Feb. 3, 2012, which had 1.41 inches of liquid and 22.7 inches of snow)
  • Denver:  0.72 inch of precipitation and 5.5 inches of snow. Third-wettest February day on record (tie)
Wet snow in Boulder, 2/3/24
Figure 2. An unusually wet snow for midwinter plasters Boulder, Colorado, on February 3, 2024. (Image credit: Bob Henson)

Along with leaving cement-like slabs of snow that challenged shovelers, the storm also served as a reminder of the complexities of predicting snowfall in a warming climate. If temperatures had been just a touch colder, the same storm might have produced two feet of snow and little or no rain.

It’s been widely noted that even in a warming climate, some winter storms may produce more rather than less snow, as large, cold mid-latitude cyclones gather more moisture. And some storms may be abetted by warming-induced deviations in the polar jet stream, a topic still being actively researched.

At the same time, other winter storms will get pushed into the too-warm-for-snow range.

Diagram showing enhanced moisture and its impact on snow and rain production.
Figure 3. Peak evaporation from lakes and oceans increases as the climate warms, allowing winter storms to potentially entrain more water vapor and produce more snow — but only up to a point, beyond which they become heavier rainstorms rather than snowstorms. (Image credit: Climate Central)

Overall, according to Climate Central, an independent group of scientists and communicators, most of the United States has seen a decrease in spring and fall snow since 1970. Winter snowfall has generally increased in the north-central states and decreased toward the south. Almost two-thirds of U.S. stations have seen a decrease in full-season snow but with many local variations (see Fig. 4 below).

Globally, an analysis led by climatologist Brian Brettschneider found that total snowfall has decreased by almost 3% since 1973.

Watch: Video: Goodbye, snow?

Maps showing where U.S. snow has increased and decreased since 1970
Figure 4. Since 1970, snowfall has increased at 36% of U.S. stations (left) but decreased at 64% of stations (right). (Image credit: Climate Central)

A fluffy but flaky storm

A few days after the cold Colorado rain, it was the Northeast’s turn for a weird winter storm, this time on the dry end of the spectrum. Forecasters had originally projected a fairly broad zone with several inches of snow that would extend from northern Virginia into New Jersey and lighter amounts into the New York City area and southern New England.

By the day of the storm’s onset, February 16, high-resolution short-term models predicted that the snow would end up focused in narrow east-west bands, leading to sharp contrasts between “winners” and “losers”. That’s basically what happened, but the results were even more dramatic than expected.

Along these narrow bands, conditions were ideal for building the type of fluffy snowflakes called dendrites, which tend to produce the largest accumulations. At some locations, the snow-to-liquid-water ratio exceeded 20:1, yielding more than twice as much snow for a given amount of moisture compared to the norm along the Northeast’s I-95 corridor (see Figure 1 above).

The main northeastern snow band extended from southeast Pennsylvania across New Jersey into the New York City area, leading to some remarkable contrasts. Some sites near Allentown, Pennsylvania, notched more than a foot of snow, and eight- to 10-inch amounts stretched across northeast New Jersey, while locations just 20 miles away as the crow flies picked up four inches or less.

In and around New York City, amounts ranged from 3.2 inches at Central Park and 4.2 inches at JFK Airport to eight to 10 inches across southernmost Brooklyn and parts of Staten Island. While forecasters correctly pegged the heavier-to-the-south citywide gradient, there was some grumbling over the higher-end totals, including a Twitter/X post from the city’s sanitation department.

After expecting one to five inches, long-suffering Washington, D.C., ended up with less than two inches across most of the metro area. A mere 0.1 inch fell at Reagan Washington National Airport, the city’s official observing site.

Veteran meteorologist Jason Samenow, who writes for Capital Weather Gang at the Washington Post, wrote a striking apologia, “Snow forecast for D.C. area was a bust. Here’s why.” Samenow asserted it was a tough forecast for the D.C. area, one that could have been delivered with more emphasis on uncertainty but also one that highlights the limits that still plague snow prediction: 

The National Weather Service and television meteorologists predicted comparable amounts or even more than we did. But, once in a while, despite all of our best efforts, we’re just going to miss the mark because the tools we have aren’t quite good enough.

As a whole, the public is much less forgiving of busted snow forecasts than busted rain forecasts. It’s easy to see why. Most people wouldn’t notice if a rainy day produced a half inch versus an inch of steady rain. But with a 10:1 ratio, that would be the difference between five inches and 10 inches of snow — which is noticeable indeed.

Snow season plays catch-up in California

Out West, the snow season has hewed closely to the El Niño playbook: dry to the north, wet to the south. Much of the Pacific Northwest and Northern Rockies had less than 70% of average snow water equivalent (the amount of water held in snowpack) as of February 24.

Farther south, a bountiful January and February have pushed California’s crucial Sierra watershed toward the 80% range for the date, with another cold storm expected this week.

And most of the Southwest U.S. is running near- to well-above-average in snowpack moisture — always a good sign in a region that’s been dominated by megadrought for almost 25 years.

Jeff Masters contributed to this post.





Source link