Sleet is often the bane of winter weather forecasting in our region. For snow lovers, it greatly cuts down snow accumulation totals. For everyone else, heavy sleet accumulations require some chiseling of car windshields, and the stuff is nearly impossible to shovel without a metal implement. It is also downright hazardous to driving.

On the days leading up to Thursday’s winter “mess,” Capital Weather Gang forecasters — and many others in our region — struggled with trying to predict relative amounts of plain snow vs. sleet. Models were waffling on the proportions, even one to two days before the event. Not until Wednesday evening, just hours before the storm, did our best models hint at more of an icy vs. snowy solution.

Whether cloud layers generate snow or sleet depends critically on the layering of temperature in and below those clouds. This is shown in the graphic below. Let’s start with snow being made in cloud layers above 5,000 feet, in subfreezing air. A wedge of milder air intrudes off the warm Atlantic, in the lower layer of the cloud.

Flakes passing through this elevated mild air partially melt into icy raindrops. But this is only half the story. Below the warm wedge is a layer of subfreezing, Arctic air trapped against the Appalachians. This layer can be several thousand feet deep, due to a process known as cold air damming. The raindrops freeze into clear grains of ice. Those ice grains, called sleet, pile up into an opaque mass of crunchy ice on surfaces.

In the above diagram, the east winds ushering in mild air aloft are shown blowing off the ocean, but milder air can arrive from a multitude of directions. Weather balloon data from 7 a.m. Thursday showed these winds coming from the southwest, with the flow overrunning a chilly Arctic flow from the north, closer to the ground.

So how does the warm layer that causes sleet manifest itself? Let’s examine temperature data, shown below, from a weather balloon that sampled the atmosphere at Dulles International Airport on Wednesday evening, hours before the precipitation began.

The red line shows the temperature measured by the balloon. It’s exactly 32 degrees at the surface, but then drops rapidly in the lowest several thousand feet of ascent labeled layer “B” representing the cold air wedge.

Critically, within the layer labeled “A” around 5,000-6,000 feet, the balloon enters an overriding wedge of milder air, where the temperature edges above freezing.

This small warm layer was a warning to forecasters in the hours before precipitation began that the atmosphere was hinting at a sleet profile. Because of this, Capital Weather Gang forecasters wrote that there was a chance that the balance of the precipitation would be sleet in our 9:30 p.m. update on Wednesday, the evening before the storm. We also said there was a chance the precipitation could still end up as snow, because sometimes heavy precipitation can help cool the column of air to freezing levels. However, while some flakes did mix in at times, the effect wasn’t strong enough to convert the sleet to snow.

All models, in general, seem to have difficulty properly simulating the arrival of a “warm plume” above the colder air mass. Sometimes this layer can be quite thin, and an error in the simulated temperature of just one or two degrees can mean the difference between a cement-like coating of sleet and a blanket of fluffy snow.

In the big picture, we can see how weather systems helped generate this warm layer. The next figure shows the morning surface weather map, with the pressure lines clearly delineating a ridge of cold, high pressure entrenched east of the Appalachians and extending from Long Island to southern Georgia. The core of Arctic air was limited to the lowest few thousand feet, again corresponding with the letter “B.”

Meanwhile, the counterclockwise circulation generated by a zone of low pressure to the west, over Tennessee, generated winds from the southwest about 5,000 feet above the ground, offering just enough mild air such that sleet would be the dominant precipitation form.

And sleet it did, at impressive rates, for several hours. Beginning late Wednesday night, some short-range models predicted that sleet would be the predominant form, after an initial burst of snow.

The following figure is a radar scan, taken from midmorning Thursday. In what appears to be a summertime thunderstorm parked over northeastern Maryland, those intense red and orange tones attest to very heavy sleet. To a radar, ice particles are highly reflective of microwave energy and a large concentration of small targets such as sleet grains will light up the radar and produce very large energy returns.

For a few hours Thursday morning, dynamical processes in the atmosphere contributed to high sleet rates. A pocket of spin energy in mid-levels of the atmosphere, which induces air ahead of it to rise strongly, was transiting the Mid-Atlantic. And a core of super-fast wind in the jet stream, called a jet streak, was also perfectly situated to further stoke upward motion.

The final figure shows the jet stream pattern at high levels, during the phase of the most intense sleet on Thursday morning.