Convective storms are the primary cause of floods and erosion from areas burned by wildfire. As part of my research as a hydrologist, I have tried to learn more about them. How does outflow from a convective area cause the development of convective storms in another area? For example, I received this message on July 15, 2016, from the Urban Drainage and Flood Control District (UDFCD) Flood Prediction Center, which serves the Denver, Colorado, metropolitan area.
A strong outflow boundary from convection over Nebraska earlier today has reached northeastern Colorado, with greater moisture existing behind the boundary. Thunderstorms have been triggered along this boundary … with additional thunderstorms possible this afternoon.
I don't understand how outflow from convective activity in Nebraska (more than 100 miles northeast of Denver) could affect storm development in the Denver-Boulder area, when the predominant upper level flow is from west to east.
John A. Moody
You are correct that the upper level flow was from the west on the day of this event. In the mid-troposphere (500 millibars, about 19,000 feet altitude), the flow was from the west-northwest at 25 mph. In the high troposphere (200 millibars, about 42,000 feet altitude), it was from the same direction at about 75 mph. One might expect that the brisk flow aloft would push any thunderstorm eastward, but on this day the air flow that caused the thunderstorms to form was in the lower troposphere and was moving toward the southwest.
Here is the sequence of events.
A large cluster of showers and thunderstorms persisted in central and southwest Nebraska for more than five hours the morning of July 15. The reflectivity image from the WSR-88D radar (KLNX) that serves western Nebraska is shown in Figure 1 for 7:23 a.m. CDT and is typical of the coverage and intensity of echoes throughout the morning. The white borders separate Nebraska from South Dakota at the top of the image and from Colorado and Kansas at the bottom. Interstate 80, in red, curves across southern Nebraska.
A radar reflectivity map for 7:23 a.m. CDT, July 15, 2016, from the WSR-88D radar, whose beam is pointed 0.5 degree above the horizontal. The radar, identified with the label KLNX, is located north of North Platte, Nebraska. Note the widespread shower and thundershower activity, which continued more than four hours after this map time. Interstate 80 (in red) curves gently across southern Nebraska. The color bar at left represents values of reflectivity factor in logarithmic units of decibels (dBZ). Higher numbers indicate stronger echoes and (usually) more intense precipitation.
Figure 2 is an estimate of cumulative rainfall from 7:00 p.m. CDT July 14 until 1:00 p.m. CDT July 15 from the KLNX radar. Because this radar has dual polarization capability, the estimate is probably not far off the mark. Rainfall was quite substantial, with areas shaded in red receiving over two inches, most of it falling on the morning of July 15. Evaporation produced a large volume of rain-cooled air. Being cooler and denser than air where rain didn't fall, this air began to spread out laterally, its leading edge marked by a gust front.
A map of cumulative precipitation between 7:00 p.m. July 14 and 1:00 p.m. July 15, 2016. The color bar at left indicates precipitation amounts. This image was derived from a time history of reflectivity data captured by the KLNX WSR-88D radar north of North Platte, Nebraska. Note that extensive areas received more than two inches of rain.
As the gust front advanced southwestward into Colorado, it was first detectable on the KFTG WSR-88D radar that serves most of northeastern Colorado, including the Denver metropolitan area, at 12:03 p.m. MDT. Figure 3 shows two disconnected gust fronts (white arrows), both generated by precipitation in Nebraska—the southern one passing through Akron, Colorado, and the other approaching Fort Morgan, Colorado.
A radar reflectivity image for 12:03 p.m. MDT, July 15, 2016, from the KFTG WSR-88 radar east of Denver, Colorado. Beam elevation is 0.5 degree, nearly horizontal. The radar is located at the red dot along Interstate 70, the east-west highway that curves across the center of the map. The major north-south highway is Interstate 25. The highway that runs northeast from Denver is Intertstate 76. From north to south, the Front Range foothills lie just to the west of Fort Collins, Boulder, Golden, Castle Rock, and Colorado Springs, Colorado. A prominent ridge runs east from the foothills through Castle Rock and Kiowa, Colorado. It rises 1,000–2,000 feet higher than the elevations of Denver and Colorado Springs. County boundaries are in green. The color bar for reflectivity factor in decibels (dBZ) is at left. Clear air radar returns (blue) cover the northeast Colorado plains. Note the two disconnected gust fronts (white arrows) just coming within radar range.
Under normal conditions, the radar beam bends downward toward the earth's surface with a curvature a little less than that of the spherical earth itself. Thus the lowest beam of the radar, pointed 0.5 degree up from the horizontal, finds itself at increasing elevations above the earth's surface with increasing distance from the radar. At a distance of about 75 miles (from the radar to Akron), the 0.5-degree beam would be about 7,000 feet above the surface, in part because the land surface slopes gently downward toward the east on the Colorado plains. This implies that the depth of the rain-cooled air was more than 7,000 feet when it was first detected.
Same as Figure 3 but for 1:07 p.m. White arrows show the gust front, advancing southwestward toward the Denver, Colorado, metropolitan area. Note showers breaking out east and northeast of Colorado Springs, Colorado.
Figure 4 shows the discontinuous gust front (white arrows), nearly 100 miles long, advancing toward the Denver metro area at 1:07 p.m. There is no precipitation in northeast Colorado. What, then, makes the gust front visible in clear air? Like a cold front, the dense, rain-cooled air is colliding with, and nosing under, warmer, less dense air. In midsummer, insect populations are high on the Colorado prairie. The convergence and uplift provided by the gust front concentrates the bugs and lifts them several thousand feet, where birds are drawn to feed on them. Bugs and birds alike are visible as targets just as raindrops are. Blowing dust may also have contributed to the gust-front echo. A third contribution is that the gust front creates a strong enough gradient in refractive index that a portion of the radar beam energy is reflected back to the radar, generating a small signal in reflectivity. All three of these factors can contribute to the reflectivity signature associated with this and other gust fronts.
Same as Figure 3 but for 1:58 p.m. The long, sinuous gust front is nearing the foothills west of I-25 by Fort Collins. It stretches southeastward almost to Hugo, Colorado. The showers that were east and northeast of Colorado Springs on Figure 4 have generated a gust front that shows as a distinct arc southeast of Kiowa. This gust front is moving east.
By 1:58 p.m., the two segments of the gust front along its northern flank have merged (Figure 5), and the KFTG radar has a better view of it because the 0.5-degree beam height at the gust front is now closer to the ground. The northern end of the gust front has just crossed Interstate 25 (red north-south highway) south of Fort Collins, Colorado. On the northeast Colorado plains, surface reports show northeast to east winds of 20–25 mph behind the gust front and, within one hour of passage, a temperature drop of 10°F, a dew point rise of 10°F, and a sky clouding rapidly with stratocumulus, similar to what happens with passage of a cold front.
By 2:49 p.m., the gust front has intersected the foothills from Fort Collins to near Boulder, Colorado, and scattered showers are forming there in response. See Figure 6. The showers near Central City, Colorado, north of Golden, Colorado, and farther south near Castle Rock and Kiowa, Colorado, are not associated with the gust front but rather would probably have formed anyway, the result of daytime heating over elevated terrain. In particular, the ridge of high ground that extends east from the foothills just south of Castle Rock is known to be a convective hotspot, where the earliest showers often form on summer afternoons.
Same as Figure 3 but for 2:49 p.m. Showers are forming from west of Fort Collins, Colorado, to Golden, Colorado, as the northern flank of the gust front pushes into the foothills. The southern flank of the same gust front, pushing west-southwest, has collided with the much smaller gust front moving east from Colorado Springs and noted in Figure 5.
By 3:18 p.m., the gust front has just passed the KFTG radar site and still retains its identity (Figure 7). Moderate thundershowers have formed between Boulder and Broomfield, Colorado, near where the gust front intersects the foothills. Animation of five-minute radar images shows that these storms generate a secondary gust front which spreads mostly southward in an arc toward Denver.
Same as Figure 3 but for 3:18 p.m. Moderate thunderstorms have formed between Boulder and Broomfield, Colorado, near where the gust front currently intersects with the foothills. These thunderstorms send a gust front (not shown) toward Denver, Colorado, during the next 90 minutes.
More than an hour and a half later, by 4:59 p.m. (Figure 8), the gust front has lost its identity, but extrapolation of its position places it north of Kiowa, where it seems to have invigorated the thunderstorm there. A strong thunderstorm has formed over central Denver. Not especially large, it nonetheless produces one-inch hail. It's possible, though hard to prove, that this storm formed from a collision between the large-scale gust front from Nebraska and the lesser gust front originating from storms that formed just south of Boulder shown in Figure 7.
Same as Figure 3 but for 4:59 p.m. The large gust front from Nebraska has lost its identity. The smaller gust front spawned by storms between Boulder, Colorado, and Broomfield, Colorado, in Figure 7 has moved south and southeast and may be responsible for the three thunderstorms from west of Golden, Colorado, to central Denver, Colorado. The storm over Denver is strong, producing one-inch hail. The thunderstorm that has blossomed near Kiowa, Colorado, is near the extrapolated position of the major gust front seen on previous figures. Like most of the thunderstorms on the afternoon of July 15, this one formed over higher ground with assistance from a gust front.
Thunderstorms on the afternoon of July 15, 2016, were widely scattered and mostly confined to elevated terrain, but why did they form when and where they did?
Every thunderstorm has three ingredients: moisture, potential instability, and a triggering mechanism.
Without water vapor in the atmosphere, there can be no clouds, no storms. The gust front arriving from Nebraska brought significantly greater moisture to the Front Range Urban Corridor and acted to increase the potential instability of the atmosphere.
To explain potential instability, it's necessary to refer to a sounding diagram, which plots measurements made by a rising balloon carrying temperature and moisture sensors. I explained the Stüve diagram in the March/April 2016 installment (pp. 38–41) of this column in Weatherwise, and so I will avoid the details here. Please refer to the Stüve diagrams in Figures 9 and 10.
The red curves show how the temperature and dew point vary when a specified volume of air (called a parcel) is lifted forcibly from near the surface. It is assumed that the parcel does not mix with its surroundings, which is a reasonable approximation, and that rain falling through the parcel does not slow it down. As indicated by the right-hand red curve, the parcel cools as it rises, and, when the temperature and dew point become equal, condensation occurs, forming a cloud. From this point on, the parcel temperature and dew point are the same. As long as the air parcel remains cooler (and therefore denser) than its surroundings (as long as the red curve lies to the left of the environmental temperature curve on the Stüve diagram), the lifting must be forced. Otherwise the parcel would sink back to its starting point.
Condensation releases energy into the parcel, raising its temperature and giving it the opportunity to become warmer than the air outside the cloud. If and when this happens, the parcel becomes buoyant at what is called the level of free convection (LFC). From this point on, the in-cloud parcel can rise much like a hot air balloon, as long as it remains warmer than the air outside the cloud (as long as the red curve lies to the right of the environmental temperature curve).
Atmospheric instability is potential until air parcels reach their LFC. The energy released within the parcel by condensation is necessary, but not always sufficient, to bring a parcel to its LFC. Unless this happens, the instability remains potential (it is not realized), and no thunderstorm will form.
Sometimes you can estimate where the LFC is by looking at a growing thunderhead. Below the LFC, the walls of the cloud are relatively smooth, where the air is being forced upward through the condensation level. At and above the LFC, the walls of the cloud bulge and swell like a cauliflower. Parcel buoyancy causes this change in appearance.
The morning sounding on July 15 from Denver (Figure 9) is fairly dry: the spread between temperature and dewpoint is substantial, and the precipitable water, which can be computed from the sounding data, is 0.52 inches. (That's the depth of a puddle that would result if all the moisture in a vertical column of the atmosphere condensed and fell to the ground as rain.) There is very little potential instability, in that the red parcel curve lies mostly to the left of the environmental temperature sounding.
A Stüve diagram for Denver, Colorado, valid at 1200 UTC (6:00 a.m. MDT) July 15, 2016. The vertical axis is pressure in millibars; the horizontal axis is temperature in °C. The two bold black curves trace the vertical variation of temperature (right) and dew point (left) as measured by a sounding balloon. Wind speed and direction, derived from balloon tracking data, are plotted to the right of the diagram. Orientation of the wind staff gives direction; small barbs represent five knots, long barbs 10 knots, and flags 50 knots. The red curves show the variation in temperature and dew point that an air parcel would experience when lifted from near the surface to above 200 mb. Cloud base is where the temperature and dew point lines intersect (low on the chart). The level of free convection (LFC) is where the curve for the parcel trajectory first passes to the right of the environmental temperature curve. On this sounding, the LFC is near 500 mb, but the parcel becomes only slightly warmer than its surroundings and doesn't stay warmer for long.
The evening sounding (Figure 10) has moistened considerably: the temperature and dewpoint are much closer together, and the precipitable water is 1.24 inches, more than double the morning value. Moistening through the depth of the troposphere has two causes: the influx of low-level moisture from the northeast and the pumping of that moisture vertically. The parcel trajectory now lies to the right of the plot of environmental temperature, from 720 to 235 millibars, despite significant cooling of the lower troposphere during the afternoon.
As in Figure 9 except for 0000 UTC, July 16 (6:00 p.m. MDT, July 15), 2016. The arrival of the gust front from Nebraska has moistened the lower troposphere considerably since the morning sounding (Figure 9), and thunderstorm updrafts have pumped the low-level moisture throughout most of the troposphere. (The spread between temperature and dew point is only a few degrees at most levels.) Because of copious low-level moisture, the parcel trajectory curve now lies to the right of the environmental temperature curve over an extended range of altitudes, making it possible to release the potential instability and build thunderstorms.
The effect of the influx of low-level moisture on potential instability can hardly be overemphasized. Increasing the low-level dew point has the effect of shifting the entire parcel trajectory to the right with respect to the environmental temperature sounding, thereby increasing the positive buoyancy of the parcel above the LFC. That, in turn, increases the potential instability.
Air must be forced upward to its LFC. If there is no potential instability, the LFC does not exist, and no amount of forced lifting will generate a thunderstorm. Forced lifting is caused when air encounters a topographical barrier or by low-level convergence.
Convergence is the coming together of different airstreams. One example is air of higher density plowing into air of lower density and forcing the latter upward. The gust front generated by rain-cooled air in Nebraska does just that, as shown on the 0000 GMT sounding (Figure 10). Following the gust front passage, the lower tropospheric winds have shifted into the northeast, whereas they were from the west on the morning sounding (Figure 9). But here is the hard-to-answer question: will the convergence along the gust front be enough to raise air near the surface to its LFC? If not, will collision of the gust front with the Front Range foothills do the job? In retrospect, it appears that upslope flow along the foothills and south of Denver was the deciding factor, though the storm over central Denver appears to have been triggered by a gust front from the storm that earlier developed near Boulder.
Gust fronts generated by thunderstorms pose a major forecast challenge. If the forecast of occurrence, location, or timing of the initial thunderstorms is inaccurate, the gust front can't be properly anticipated. Once formed, whether the gust front will be able to raise parcels of low-level air to their LFC is an even more vexing question. In the case just reviewed, it is likely that the additional lift provided by flow against the foothills was just enough to bring parcels to their LFC, because very few storms fired over the relatively level plains.
Here are the relevant questions a forecaster must answer in order to successfully predict thunderstorms spawned by gust fronts: (1) Where will the initial thunderstorms develop? (2) Will they produce enough evaporatively cooled air to generate a gust front? (3) Where will the gust front go? (4) Will it interact with rising terrain, an existing convergence zone, or another gust front? (5) Will rising parcels reach their LFC? Very detailed observations of temperature and moisture in the lowest few thousand feet are required to answer the last question, and, in most cases, such observations are not yet available.
Readers for whom the discussion of the Stüve stability diagram is too much might want to consult the discussion of “Vertical Air Movements,” on pages 81–85 of The AMS Weather Book. The Ultimate Guide to America's Weather, published jointly in 2009 by the American Meteorological Society and the University of Chicago Press. Author Jack Williams manages to convey the same ideas but without the complicated diagrams.
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 email@example.com, or by mail in care of Weatherwise, Taylor & Francis, 530 Walnut St., Suite 850, Philadelphia, PA 19106.
Paul T. Schlatter, at the National Weather Service's Denver, Colorado, office, procured the radar images. John C. Osborn, at NOAA's Earth System Research Lab in Boulder, Colorado, annotated the figures.