I live in Baltimore, Maryland. On the evening of Friday, April 10, 2015, we experienced frequent thunder for nearly an hour. Though the storms weren't particularly strong, they occurred with an air temperature of 42°F. I have experienced “thunder-snow” a few times and even “thunder-sleet” once during winter, but in those cases the thunder lasted only a few minutes. How could these thunderstorms persist so long in such cool air?
I'll give a short answer first, then a longer, more detailed answer. A weak cold front traversed the Baltimore area on the evening in question. Though no atmospheric soundings are available for Baltimore, it is quite likely that a shallow layer of cold air overspread the area after frontal passage. Above this layer, the air was warmer, nearly saturated (relative humidity close to 100%), and conditionally unstable. The latter condition means that updrafts initiated in the air mass riding over the top of the low-level cold air could support deep convective clouds in the same way that a modest thunderstorm might grow from an updraft forming very close to the ground on a summer day.
Relatively warm, moist air, riding up over a sloping cold frontal surface can be lifted sufficiently to initiate convection, as illustrated in Figure 1. If thunderstorm cells line up along the wind direction aloft, multiple cells traverse the same location. Meteorologists call this training because the cells act like cars on a train track. They all cross the same spot.
I could not find enough weather data for the right time of day to prove my statements about the Baltimore weather, so I took the next best step. I found a clear-cut example of what is called elevated convection—convection that has its roots not at the ground but rather in an air mass above the ground. This example, which occurred in Oklahoma on March 8, 2012, closely matches what you observed.
Figure 2 is a plot of surface data centered on Oklahoma. A cold front stretches from Missouri to south of the Texas Panhandle. Immediately north of the front, temperatures are in the 40s, while south of the front the air is warm and humid (temperature in the low to mid-60s and dew points only a degree or two lower). Note the symbols for rain on both sides of the front and a few symbols for thunderstorms, especially in southwest Oklahoma. The colder, denser air is plowing into warm, moist, and less dense air, which is forced upward over the frontal surface.
Figure 3, an analysis of conditions on the 850-mb pressure surface (roughly 4,900 feet above sea level), illustrates this clearly. In Figure 2, Oklahoma City, at the very center of the map and the center of Oklahoma, registers light rain, a brisk north wind of 25 knots, and a temperature of 42°F. At 850 mb, the temperature at Norman, Oklahoma, just 20 miles south, is 11°C (52°F), 10°F warmer than at the surface, and the wind is from the southwest at 20 knots. Clearly, the warm, moist air is riding up over the cold air near the surface.
Note the temperature contours (isotherms) drawn for every 2°C: dashed red for temperatures above 0°C, and dashed blue for temperatures 0°C and below. The bunching of isotherms not far northwest of Norman indicates the position of the cold front at 850 mb. The green contours are drawn wherever there are eight grams or more of water vapor within each kilogram of dry air. The plot shows a strong southerly wind over Texas and southern Oklahoma that is importing copious moisture into the frontal zone. This air is conditionally unstable, as we shall see.
Figure 4 is the most complicated chart you will ever see in this column. It is a plot of measurements made by a rawinsonde balloon launched at Norman, Oklahoma, and valid on March 8, 2012, at 1200 GMT, the same time as in Figures 2 and 3. Despite its complexity, I show this chart because it illustrates well the concept of elevated convection. It is called a Stüve diagram. It is not the same as the Skew-T/Log p chart that most meteorologists use, but it is somewhat easier to understand.
First, consider the rectangular grid. The vertical lines are for temperature or dew point (°C); they are spaced at 5°C intervals. The horizontal lines are for pressure (in millibars). They are spaced at 50-mb intervals. The pressure intervals widen as pressure decreases, so that the vertical scale on this diagram is approximately linear. (If you move an inch upward on this chart, the change in altitude will be about the same, regardless of pressure.) Nominal altitudes in meters are plotted just to the right of the pressure labels along the left side of the chart.
What measurements, obtained from instruments suspended from the rising balloon, are plotted on this chart? Winds are plotted along the right side of the chart. The location of the foot of the wind staff indicates the pressure where the wind was measured. The orientation of the staff indicates the wind direction. Read the direction as you would a compass. Up is north, right is east, down is south, and left is west. The staff points into the wind. The barbs at the top of the staff indicate wind speed: a short barb represents five knots, a long barb 10 knots, and a flag 50 knots. As an example, the wind at 300 mb is from the WSW; the speed is 75 knots (one flag + two long barbs + one short barb).
Temperature and dew point appear as the wiggly, bold, black lines—temperature on the right, dew point on the left. These indicate how the temperature and moisture content of the air vary with pressure (equivalently, altitude). The greater the spread between temperature and dew point, the less the relative humidity, and the less likely are clouds. This sounding is “moist” because the temperature–dew point spread is nowhere greater than 10°C below 300 mb. In fact, the temperature and dew point are the same below 750 mb, indicating saturated air and clouds.
Note the frontal inversion just below 900 mb. It is 7.5°C (13.5°F) warmer at the top of the inversion than just below it. The inversion marks the top of the shallow cold air mass. Note the wind shift from north at 15 knots close to the ground to south-southwest at 10 knots just above 900 mb.
As one proceeds north behind the cold front shown in Figure 2, the depth of the cold air increases and so does the height of the frontal inversion. The warm, moist air must ascend this sloping frontal boundary because of the difference in density between the two air masses. This lifting is enough to initiate spotty convection, as demonstrated next.
So far I have discussed the environmental measurements plotted on the Stüve diagram. Now it is time to discuss what happens to air at a given temperature when it is lifted from a given pressure. The straight, diagonal, green lines on the chart are called dry adiabats. If an unsaturated volume of air (relative humidity less than 100%) is lifted from a given temperature and pressure, its properties will follow a dry adiabat. For example, note the dry adiabat that passes through 0°C and 1000 mb. If a volume of air initially at 0°C is lifted from 1000 mb, its trajectory follows the green line. If no condensation occurs, this volume of air will have cooled to –50°C by the time it reaches 500 mb. The same dry adiabat (green line) connects the two end points.
Because all air includes some water vapor, the cooling that accompanies lifting may eventually initiate condensation. Once condensation begins, the volume of air no longer cools at the dry adiabatic rate because condensation releases energy into the volume. The rising volume will now follow a moist adiabat and have a cooling rate less than it did before it became saturated. Moist adiabats are given by the gently curved blue lines in the Stüve diagram. At low temperatures, the moist adiabats look a lot like dry adiabats because so little moisture is left to condense.
I will not further complicate matters by discussing the purple lines on the Stüve diagram because they are not relevant here.
Two process curves appear on the diagram. These curves show what happens to air volumes that are forced upward from starting points at 750 mb and the surface, respectively. From the plotted soundings, it is clear that both volumes are saturated (the temperature and dew point are the same). A volume of air initially saturated remains so during ascent. The red curve shows what happens to a saturated volume when lifted from 750 mb. Following a moist adiabat, it immediately becomes warmer than its surroundings (because the measured temperature lies to the left of the red curve). This volume of air is thus buoyant and will rise of its own accord, somewhat like a hot-air balloon, until it is no longer warmer than its surroundings, which happens at about 250 mb. A cloud growing to 250 mb is easily tall enough to generate thunder and lightning. The trick is to force the air at 750 mb to rise in the first place. As we shall see next, ascent of warm, moist air over the cold frontal boundary is, at least in some locations near and behind the front, sufficient to initiate convection.
The other process curve starts from the surface. Like the red curve, it follows a moist adiabat, the one colored bright blue. This air volume remains negatively buoyant (denser than its surroundings), and consequently has no chance to build a convective cloud.
Finally, Figure 5 is an image of reflectivity from the NWS radar at Twin Lakes, Oklahoma, near Norman. It is valid at the same time as previous figures. Radar echoes colored green represent light precipitation; those colored yellow, orange, or red indicate heavier precipitation with thunder and lightning likely in the more intense cells. The white curve indicates the position of the surface cold front. Thunderstorms lie on both sides of the front, but those on the north side are elevated (have their roots above the cold frontal inversion, not at the surface). As the Norman sounding indicates in Figure 4, the steering winds in the mid- and high troposphere are generally from the southwest. Thus, thunderstorms oriented along any northeast–southwest line are likely to trail each other. If they are closely spaced, thunder could continue at favored locations on the ground until the entire group has passed. This could happen in southwest Oklahoma or close to the frontal boundary southeast of Oklahoma City (where the major highways converge).
To summarize, thunderstorms that occur when the surface temperature is close to or below freezing often (but not always) result from an overrunning situation, when relatively warm, moist air rides up over a surface-based cold air mass. If the slope of the frontal boundary is relatively steep, say, rising from the surface to an altitude of several thousand feet above ground 100 miles behind the leading edge of the cold air, and if the overrunning air is conditionally unstable, then elevated thunderstorms may form over the cold air mass. If thunderstorm cells line up along the steering currents in the mid- and high troposphere, then echo “training” can occur and rumbles of thunder may be prolonged at a given location.
Weatherwise Contributing Editor THOMAS W. SCHLATTER is a retired meteorologist and volunteer at NOAA's Earth System Research Laboratory in Boulder, Colorado. Submit queries to the author at firstname.lastname@example.org, or by mail in care of Weatherwise, Taylor & Francis, 530 Walnut Street, Suite 850, Philadelphia, PA 19106.