In this article we discuss the meteorology behind our region’s ice storm, and how the Mid Atlantic’s unique geography provides an ideal setup for these highly disruptive weather events.

More than just a nuisance

A heavy snowfall has the ability to polarize people into two camps – the snow lovers and snow haters.  But I’ve never met a soul who cheered on a good ice storm, for all the misery these events bring: Threats to our personal safety, travel nightmares, damage to trees and extended power outages.

The storm that unfolded during the early morning hours of February 5 was well-forecast, several days in advance.  The CWG team (along with the NWS) successfully predicted the large gradient in ice accumulation, ranging from a thin coating south and east of D.C., to greater than half an inch along the Mason-Dixon Line.

South-central and southeast Pennsylvania was ground zero for the heaviest ice accumulation, with nearly 500,000 power outages in that region alone.   But north-central Maryland was also crippled by outages, as the figure 1 below illustrates.

Figure 1: A county-by-county tally of utility outages across Maryland as of 12:30 pm, February 5.  (Maryland Emergency Management Agency)

A re-examination of the map at 5:30 p.m. (not shown) showed largely the same set of colors – meaning the affected counties were bracing for a multi-day power restoration.

To understand why north-central Maryland was so hard hit, it’s necessary to briefly examine the relationship between our region’s terrain and air temperature.  If you want to blame something for these ice storms, it’s the Appalachians.

The Appalachians be dammed

Those of you who are CWG regulars may have heard The Gang discuss Appalachian cold air damming (CAD).   To any meteorologist and forecaster working in this region, CAD is “bread and butter” when it comes to understanding our cool season’s climate.

Climatological studies of CAD events reveal that our region experiences 2-3 of these episodes, on average, each month from December through March.   Not all CAD events are necessarily intense, nor associated with precipitation.  Some bring locally chilly air, others a dreary deck of stratus overcast.  But if there’s an ice storm in the Mid Atlantic, it’s almost guaranteed that CAD is part of the setup.

Cold air damming events have already contributed to three snow and/or ice episodes this season (December 8-9, 2013;  January 10, 2014;  and January 21, 2014).

Figure 2 shows what happens during CAD.   The Appalachians, which range in elevation from 2,000 ft in Maryland to more than 4,000 ft in central Virginia, create an imposing obstacle to cold air filtering down from the north.

Figure 2: Airflows involved with CAD and overrunning leading to precipitation along the eastern side of the Appalachians.  Adapted from Bell and Bosart (1988).

The dense air arriving from New England (pumped southward by high pressure parked over that region) cannot surmount this geological rampart.  Instead, a cold air current frequently develops along the eastern slopes, directed from the north.  Such a “boundary current” (low-level wind maximum – blue arrow) builds a cold, high pressure dome of air only a few thousand feet thick.

If a weather disturbance approaches from the west or south, it pulls in a mild, easterly flow off the Atlantic.  This current (red arrow in Figure 2) overruns the cold dome.  At a minimum, the uplift generates stratiform cloud layers.  With enough moisture and vigorous uplift, precipitation develops.

If air temperature in the cold dome is below freezing, the result is sleet or freezing rain, depending on the depth of the cold air.

Figure 3 presents another view of CAD, seen from the vantage of a west-east slice through the atmosphere, crossing the Appalachians.  The blue shaded regions indicate the shallow dome of cold air pushed up against the east slopes of the mountains.

Figure 3: Vertical structure of CAD showing a 3,000 ft thick cold dome of air wedged against the eastern slopes of the Appalachians.    Note the northeasterly winds (barbed arrows) feeding cold air into the region.

The real deal on February 5

Figure 4 (below) shows a snapshot of the actual cold air wedge at 11 pm,  February 4.  The weather chart shows the pattern of air temperature (dashed lines = isotherms) at an altitude of about 2,000 feet.

Figure 4: The classic CAD thermal signature at the 2,000 ft level.

Red isotherms portray temperatures above freezing, in 1°C increments, whereas blue isotherms are those 0°C and colder.   One doesn’t need a map of the terrain to pick out the conspicuous southwest-northeast trending wedge of sub-freezing air bisecting PA-MD-VA.

The thermal gradient is incredibly tight between south central Pa. and Tidewater, Va, amounting to nearly 30°F of temperature variation.   Through the morning hours, persistent southeasterly winds attacked the cold wedge, slowly warming the Virginia and Maryland Piedmont above freezing.

By mid-late morning the precipitation had moved out while this slow warming was under way.   Surface temps reached nearly 40°F in spots close to D.C., enabling some vigorous ice melt.

Forecasting the threat to utilities

A combination of heavy ice accumulation on trees and strong wind is a sure-fire recipe for utility disaster.  These two factors can occur in various combinations, and these permutations have been captured in a new predictive tool for utility mayhem (Figure 5).

Figure 5: The Sperry-Piltz Ice Accumulation of SPIA Index, charting the likelihood of utility disruption as a function of wind speed and glaze ice accumulation.

This is the SPIA Index or Sperry-Piltz Ice Accumulation Index, and it is being tested at NWS offices at various locations such as Oklahoma.   However, predictions are routinely run for any location within the CONUS, and we at CWG were able to assess our own SPIA threat the evening of February 4:

Figure 6: SPIA Index prediction for the Mid Atlantic’s February 5-6 ice storm.

The particular forecast shown here was valid for 7 p.m. February 4 through 7 PM February 5, encompassing the entire storm.  Note the correspondence between SPIA Level 2 (orange – scattered interruptions lasting 12-24 hours) and regions of high outage density in Figure 1 (above).

While SPIA correctly identified the region of impact (from south central PA into north central MD), it seems that the threat’s magnitude was under-estimated by an entire category.  Category 3 (Red) predicts a widespread, multi-day outage.   That’s what folks along the Mason-Dixon are dealing with.  For the range of observed ice accretion across the region (0.25” to 0.5”) and actual winds (less than 10 mph), Category 2 was the predicted threat.

Why the discrepancy?  I suspect that an additional factor was not accounted for in the SPIA algorithm – namely, several inches of heavy, wet snow clinging to trees from the February 3 snow storm.    The heaviest accumulations (6-7”) were along the Mason-Dixon Line during that event.   Now, add the weight of 0.5” ice onto each branch and twig (ice has a density 90% that of liquid water).  This essentially “primed” the region for a lower threshold of widespread utility damage.