Blizzard of 2015 Brings 2 - 3' of Snow and Questions About Forecast Accuracy

The snows from the Blizzard of 2015 have finally ended over most of New England, leaving some truly historic snowfall totals. The biggest snows hit Central Massachusetts, with three feet measured at Auburn, Hudson, and Lunenburg. More than two feet of snow fell across five other states, with 33.2" at Nashua, NH; 30" at Orient, NY; 28.5" at Burrillville, RI; 31.5" at Sanford, ME; and 34.1" at Winthrop, CT. Some superlatives for stations with a long period of snow records:

Worcester, Massachusetts: All-time record snowfall (34.5")
Boston, Massachusetts: 6th heaviest snowstorm (24.4")
Providence, Rhode Island: 4th heaviest (19.1")
Portland, Maine: 4th heaviest (19.1")
Blue Hill Observatory, SE Massachusetts: 2nd heaviest (30.8")


Figure 1. A combination of the day-night band and high resolution infrared imagery from the Suomi NPP satellite shows the blizzard near peak intensity as it moves over the New York through Boston Metropolitan areas at 06:45Z (1:45 am EST) on January 27, 2015. The nighttime lights of the region are blurred by the high cloud tops associated with the most intense parts of the storm. Visit NOAA's Environmental Visualization Lab for a spectacular hi-res version of the image.

The top six snowstorms on record in Boston, Massachusetts since 1890:

1. 27.6" Feb 17-18, 2003
2. 27.1" Feb 6-7, 1978
3. 25.8" Feb 24-26, 1969
4. 25.4" Mar 31-Apr 1, 1997
5. 24.9" Feb 8-9, 2013
6. 24.6" Jan 26-28, 2015

The top seven snowstorms on record in Providence, Rhode Island, since 1905:

1. 28.6" Feb 6-7, 1978
2. 23.4" Jan 22-23, 2005
3. 22.8" Jan 7-8, 1996
4. 19.1" Jan 26-28, 2015
5. 18.9" Feb 14-16, 1962
6. 18.3" Feb 3-4, 1961
7. 18.0" Feb 8-9, 2013

The top 4 snowstorms on record in Portland, ME:

1) 31.9" Feb. 8-9, 2013
2) 27.1" Jan. 17-18, 1979
3) 25.3" Feb. 17-18, 1952
4) 23.8" Jan 26-28, 2015

The Blue Hill Observatory in Southeast Massachusetts, with snowfall records going back to 1885, recorded 30.8"--its second greatest snowstorm on record. Their top 5 snowstorms on record:

1. 38.7" Feb 24-28, 1969
2. 30.8" Jan 26-28, 2015
3. 30.3" Mar 3-5, 1960
4. 30.1" Feb 6-7, 1978
5. 30.0" Mar 31 - Apr 1, 1997

The storm's powerful winds, gusting as high as 78 mph at Nantucket Island, brought a significant storm surge to the coast. As storm surge that peaked at 4.51' at 15:18 UTC January 27 hit Nantucket Island, flooding downtown Nantucket to a depth of 3.5'. The storm surge at Boston Harbor peaked at 4.78' at 16:12 UTC January 27, and caused damage to buildings and roads all along the Southeast Massachusetts coast.

There will be no rest for snow-weary New Englanders in the near future. A clipper-type system will drop light to occasionally moderate snow across the Northeast on Friday - Saturday, followed by an intense shot of bitter cold. The extended-range models suggest a much stronger system early next week, a potential nor'easter that could deliver heavy precipitation from the mid-Atlantic to Maine. There are still major uncertainties about timing, location, and precipitation type, as this storm will carry a good bit of warm, moist air--and our experience with this week's storm (see discussion below) should give us pause before jumping too quickly on any particular model solution.

An extreme jet stream pattern brought blizzard to the Northeast, record January warmth to the West
This week's Northeast U.S. blizzard was triggered by an unusually extreme jet stream pattern, featuring a sharp ridge of high pressure along the U.S. West Coast and a deep trough of low pressure diving to the south over the Northeast United States. This configuration allowed cold air to spill out of the Arctic behind the trough to feed into the blizzard, while at the same time allowing anomalously warm air to flow northwards across the Western U.S., a contrast that’s been playing out repeatedly across the nation for more than a year. More than 40 daily record highs and 7 monthly highs were tied or broken last weekend in California, Oregon, and Washington, according to NOAA’s U.S. records website; most notably, Death Valley, California hit a remarkable 87° on Sunday, tying its record for warmest January day on record. Many more records were smashed across a broader swath of western and central states early this week. East Rapid City, SD, had its warmest January low on record Monday, dipping only to 50°F, and Denver, CO, saw its warmest temperature for any January day in 127 years, with a balmy 75°F on Tuesday (exceeded only by a 76°F reading on the same date in 1888.) Numerous stations set or tied records for all-time warmest January day on record Monday and Tuesday:

North Platte. NE (74°)
Imperial, NE (77°)
Valentine, NE (72°)
Goodland, KS (79°)
Hill City, KS (83°)
Colby, KS (82°)
Great Falls, MT (67°)

…and a number of daily records were broken by impressive margins on Tuesday, including:

Dickinson, ND: 61°F (old record 51°F)
Oklahoma City, OK: 78°F (old record 71°F)
Russell, KS: 83°F (old record 65°F)


Figure 2. An extreme jet stream patten observed at 7 am EST (12 UTC) on January 27, 2015. Color-coded wind speeds at a pressure of 300 mb (roughly 9,000 meters or 30,000 feet) show the axis of the jet stream over North America, with a large upside-down "U"-shaped ridge of high pressure over the West Coast. All-time record high temperatures for the month of January were observed over several locations in the Western U.S. underneath this ridge. The strongest winds of the jet stream (orange colors, 160 knots) were observed over the Southeast Canada, downstream from where a strong "U"-shaped trough of low pressure was anchored. The surface low pressure system associated with this trough brought the Blizzard of 2015 to the Northeast U.S. As I've discussed many times (most recently in my April 2014 post, California Drought/Polar Vortex Jet Stream Pattern Linked to Global Warming), extreme jet steam patterns like this one have increased in recent years. I wrote a story for the December 2014 issue of Scientific American called, "Is the Jet Stream Getting Weird?" which discusses how climate change could potentially be responsible for this increase in extreme jet stream patterns. Recent runs of the GFS and European models point to a highly amplified jet stream pattern continuing over North America and spreading to Europe next week, which will produce a variety of extreme weather over both continents. This image was generated from the 12 UTC January 27, 2015 run of the GFS model, and plotted using our wundermap.

The NYC forecast: Moving from diagnosis to treatment
The underwhelming snow totals in New York City are on track to become one of the most famous U.S. forecast “busts” of the 21st century. Many journalists, bloggers, and meteorologists donned their rear-view goggles on Tuesday to analyze what went wrong. Most of them, including the New York Times, correctly observed that in the big picture, this wasn’t a bust at all. NWS computer models and public outlooks captured the nor’easter as a whole--including its extreme intensity--remarkably well. The record snowfalls in eastern Massachusetts bore out the pre-storm designation of “historic.” The devil lay in the details of predicting where the storm’s west edge would be, as a dramatic gradient in snowfall totals (which forecasters did anticipate) ended up being overlaid atop the nation’s largest metro area. As discussed in our post yesterday, forecast models diverged markedly on where to place that crucial western edge. In turn, there were sharp differences among various forecast sources on what would happen in New York City, even as late as Monday evening. In a press conference with reporters on Tuesday, NWS director Louis Uccellini strongly defended his agency’s science and staff while acknowledging, “We all know in this business you’re only as good as your last forecast.”

Clearly, the outcome for New York was more uncertain than for Boston, where there was much greater model agreement on the risk of two to three feet of snow. Probabilistic tools, such as the SREF plume featured in our post yesterday, are one way that the NWS captures such distinctions. Another is the nationwide probabilistic winter weather outlooks issued by the NWS’s Weather Prediction Center twice each day. Some NWS local offices also provide experimental online graphics showing the potential best- and worst-case outcomes. “This is an important step forward,” noted Jason Samenow (Capital Weather Gang), “but these pages are difficult to find on their Web sites, and this information is not included in public forecasts.” The University of Washington’s Cliff Mass summarized the New York forecast and its implications in his blog: " . . .the NWS has to move to a much more probabilistic form of forecasting preparation and dissemination, one in which forecast uncertainties are made clear to users. The computer workstations used by NWS forecasters and NWS websites are not designed to facilitate probabilistic prediction. This needs to change."

Regardless of how well probabilities are assessed and conveyed, it’s still up to the recipients of forecasts to make critical yes/no decisions. University of Georgia’s Marshall Shepherd put it this way in a Weather Underground blog post: “There is more risk and nuance in weather forecasts than the public is interested in consuming, so it is a challenge to craft a message that gets attention, is not 'hype', yet has actionable information. We must continue to have the discussion about how to communicate uncertainty and risk effectively.” Social scientists are increasingly collaborating with NOAA on research-based strategies for improving how forecasts are communicated. NCAR’s Julie Demuth has been involved in such work on hurricanes and tornadoes. According to Demuth, though, “very, very little social science work has been done on winter weather events.” There’s also the growing problem of information overload. Weather forecasters have to sift through an enormous amount of data under intense stress and time pressure. Consumers have to reconcile what they hear about an impending storm from public and private outlets, family and friends, colleagues, and other sources, and then decide how and when to act. Analyzing this complex chain of events isn’t an overnight project, but the recent experience in New York suggests that taking the time to learn how best to convey winter storm threats could be well worth the effort.


Figure 3. 8:20 am Brooklyn, Tuesday January 27. Three days of emergency food stockpiled for this? A Brooklyn streetscape early on Tuesday morning in the wake of less-than-expected snowfall. Image credit: wunderphotographer hartKitt.

Bob Henson and Jeff Masters