Published Mon May 24, 2021
On June 3, 1981, Denver was struck by two F2-rated tornadoes causing 42 injuries as well as the destruction of nearly 100 homes in the metro area.
This is how the National Weather Service described the event:
In 1981…severe thunderstorms produced tornadoes over metro Denver. The first tornado touched down at the intersection of Alameda Ave. And Sheridan Blvd. The twister moved north along Sheridan Blvd….damaging businesses…apartment buildings…homes…and vehicles. Over ten homes were unroofed. The roof of one landed in the middle of a neighborhood park. At least 10 mobile homes were wrecked.
The tornado curved to the northeast into northwest Denver… Hopping up and down in several places. Very strong winds outside the actual funnel caused 20 to 30 thousand dollars in damage in downtown Denver. The third floor of one old building was demolished. No major injuries were reported from the tornado…although several people were hurt slightly in traffic accidents on Sheridan Blvd. in the confusion caused by the storm. Damage in Lakewood alone was estimated at 200 thousand dollars.
At the same time… The worst tornado to ever hit metro Denver struck Thornton. Coming from the same thunderstorm that spawned the Denver twister…the Thornton tornado tore a swath through the heart of the city. 87 homes were destroyed…110 others damaged at least moderately. In all…600 homes in a 100 block area sustained some damage. The twister also hit shopping centers…several restaurants…and other buildings. Seven of the 42 injured were considered serious. The storm was strong enough to snap lamp posts in half and drive a 6- inch slab of wood 2 feet into the ground. Damage was estimated at up to 50 million dollars.
Research conducted by Szoke et al. (1984) concluded the Denver Convergence-Vorticity Zone (DCVZ) was responsible for this event. A second significant Denver tornado event occurred on June 15, 1988, involving two F2-rated tornadoes and one F3-rated tornado, one of which touched down in Denver near Broadway and Evans.
This event was well documented and led to the development of the non-supercell tornado (NST) lifecycle model (Szoke* et al., 2006). Non-supercell tornadoes (NSTs) are sometimes referred to as landspouts. The processes that form NSTs differ from supercell tornadoes, which we’ll cover in more detail later in the article.
On June 18, 2013, the DCVZ produced a NST on the Denver International Airport (DIA) runway. The tornado touched down in close proximity to DIA’s Terminal Doppler Weather Radar site (TDEN), its Automated Surface Observation System (ASOS) station, and the Denver WSR-88D doppler radar (KFTG).
The NST was visible on radar. The pink arrow in the image below points to the hook echo associated with the tornado, which is located just southeast of DIA’s east terminal.
This event was well documented on social media by astonished travelers, who could clearly see the tornado from inside the terminal. Unsurprisingly, the airport’s bathrooms also double as tornado shelters.
More recently, a DCVZ event on May 26th, 2019, produced three tornadoes in Weld and Adams counties, Colorado. Typically, the DCVZ active season falls within the same climatological range as High Plains severe weather: mid-May through mid-August, with peak activity during the month of June (Smith, 2019).
However, a DCVZ event can occur any month of the year. On October 4, 2004, 11 different NSTs were observed along the DCVZ from Denver International Airport (DIA) to Weld County, causing minimal property damage. This off-season outbreak was unexpected and demonstrates the difficulty of forecasting DCVZ events.
High Plains Severe Weather and Colorado Topography
Weather balloons (radiosondes) are launched at least twice a day all over the world so weather organizations and meteorologists can gather and analyze data on conditions throughout the atmosphere. These measurements are called atmospheric soundings. A radiosonde is launched from Denver’s Central Park neighborhood twice a day, every day (sometimes more often, depending on the chance for severe weather or other phenomena).
This data can be plotted on a Skew-T diagram, which gives meteorologists quite a bit of information about the vertical profile of the atmosphere.
A typical High Plains severe weather sounding will show dew point temperatures as low as 45°F, steep lapse rates greater than 7 °C*km⁻¹, a low-level moist layer, and convective available potential energy (CAPE) less than 500 J/kg. All of this information is available on the Skew-T diagram.
DCVZ severe weather events typically initiate under these conditions with weak winds aloft. These conditions are different than what often form severe storms over the plains in the Great Plains states east of Colorado, where forcing and lift is due to synoptic scale processes — e.g. the interaction of large-scale phenomena in the atmosphere. These phenomena include low pressure systems, weather fronts, and mid-latitude cyclones, which are hundreds or thousands of miles in size.
In the absence of synoptic-scale lift in Colorado, forcing is often a result of convergent air masses in the planetary boundary layer — the lowest part of the atmosphere. In Colorado, low level forcing can include an outflow boundary (OFB), or slope drainage winds (Schreiber-Abshire & Rodi, 1991).
These are mesoscale processes, e.g. phenomena and interactions that occur on an intermediate scale of less than a few hundred miles. These irregularities, in comparison to traditional Great Plains severe weather criteria, add complexity to forecasting High Plains severe weather events.
In addition to the above processes, Colorado contains a number of prominent terrain features that form topographical boundaries which greatly influence the weather of the High Plains.
For the purpose of this article, the High Plains is defined as a gradually sloping area between Colorado’s eastern border (about 4,000ft / 1,200m above sea level) and the base of the foothills (Denver: 5,280ft / 1,609m) just east of the Front Range mountain range (highest point: Grays Peak, 14,278ft / 4,352m) in the Southern Rocky Mountains of North America.
The Rocky Mountains are the most obvious terrain feature influencing the weather of the High Plains, but another very important topographic boundary is the Palmer Divide: a ridge in central Colorado that runs east/west from the town of Limon to the foothills, which features elevations of 6,000–7,700ft (and a few hills a bit higher than that). The Palmer Divide separates Denver and Colorado Springs, which often results in both cities receiving remarkably different weather impacts from storms in both winter and summer.
Several studies (Bann et al., 2003; Schreiber-Abshire & Rodi, 1991; Wakimoto & Wilson, 1988) have found that convergence-vorticity zones form in the High Plains when ambient southerly or southeasterly flow breaches or circumnavigates a topographic barrier (the Palmer Divide) and encounters another air mass with conflicting flow (such as northwesterly flow over/down the mountains of the Front Range). The convergence-vorticity zone is defined as the boundary where these air masses meet.
The most prominent example is the Denver Convergence-Vorticity Zone (DCVZ), which forms north of the Palmer Divide and near the city of Denver, but other convergence-vorticity zones have been documented north of the Cheyenne Ridge (Wyoming) and Pine Ridge (Nebraska and South Dakota) (Bann et al., 2003).
These zones can aid storm development and thus correspond with an increase in severe storms but do not necessarily produce lift or storms on their own (Bann et al., 2003). The convective potential (the ability to produce strong storms) of a convergence-vorticity zone is often enhanced by low-level return flow after the air mass encounters a second topographical boundary (Szoke* et al., 2006).
A defining aspect of the DCVZ is that it is formed specifically through the small-scale interactions of low-level winds with Denver’s surrounding terrain. This is not to be conflated with convergence zones and boundaries that often form in the region through other interactions on a larger scale, e.g. a cold front moving through the area. If this occurs, even in the Denver area, the resultant zone of convergence and vorticity is not technically the DCVZ, and is instead a more “classic” severe storm setup that is not specific to the High Plains, unlike the DCVZ.
DCVZ Development and the 1981 Denver Tornadoes
Because the DCVZ is a result of low-level winds interacting with local terrain, it is considered a semi-stationary feature, meandering misoscale¹ distances with wind shifts.
“Textbook” High Plains severe weather synoptic conditions were observed in the 1981 case study by Szoke et al. (1984): CAPE less than 500 J/kg, weak winds aloft, and a low-level easterly or southeasterly jet advecting a moist layer across the eastern Colorado river valleys. The paper states:
Strong convergence and cyclonic vorticity are concentrated in a narrow band from just south of Denver north to [Greeley], as southeasterly flow over the plains confronts northerly and even northwesterly flow in the region of lighter winds west of the convergence zone.
In an idealized scenario, as daytime warming begins, surface dew point temperatures remain mostly unchanged. By the afternoon, mid-level southeasterly flow gradually shifts to southwesterly, strengthening the cyclonic surface circulation west of the convergence zone.
The wind shift along the area of convergence creates strong vertical shear, resulting in a line of misoscale vortices, which we will refer to as a vortex line, hours before convection gets underway.
Once the convective temperature is reached (often in the afternoon), the DCVZ is fully established. At this point, any form of weak forcing, such as an outflow boundary (OFB) from a nearby rain shower, can initiate explosive thunderstorm development along or in the vicinity of the convergence-vorticity zone (Szoke et al. 1984). This is key, as again, the convergence-vorticity zone does not initiate thunderstorm development on its own.
In the case of outflow boundaries, their associated wind gusts can be enough to initiate convection in an environment that is otherwise just stable enough to be unable to produce thunderstorms on its own.
In the summer, it is common for thunderstorms to form every afternoon in the highly unstable atmosphere above Colorado’s mountains, only to collapse and produce OFBs as they move eastward and descend in elevation to a more stable environment.
A case study Szoke et al (1984) found that convection along the DCVZ in the 1981 tornado event was initiated through this process — an OFB from benign thunderstorms descending west from the foothills. Within 20 minutes of the OFB intercepting the DCVZ, a sizable and severe thunderstorm had rapidly developed over the Denver suburb of Lakewood, producing an F2-rated tornado. This particular storm continued to strengthen and move north-northeast, producing a second F2-rated tornado in the suburb of Thornton (Szoke et al. 1984), with winds up to 157mph.
More recent studies have discovered a moisture gradient along the eastern side of the vortex line, which strengthens vertical vorticity advection along the convergence zone (Pietrycha et al., 2006). Additionally, they have discovered the existence of the vortex line along the DCVZ. These signatures can sometimes be detected with base velocity radar before any visible sign of severe weather is present (Szoke et al., 2014).
Prior to convection, merging circulations along the vortex line increase chances of tornado formation (Markowski et al., 2016). However, the presence of the vortex line itself does not produce a thunderstorm or any visible rotation. A tornado is produced only when deep moist convection (DMC) is initiated at or passes over an active DCVZ.
Convergence and Non-Supercell Tornadoes
Tornadoes which form via non-supercellular processes are formally known as Non-Supercell Tornadoes, or NSTs (Wakimoto & Wilson, 1988), which can include satellite tornadoes on the flanking line of a supercell.
For clarification, NST discussion in this article is limited to tornadoes which form beneath storms lacking a mesocyclone. The phrases “non-supercellular tornadoes” and “non-mesocyclonic tornadoes” are often used interchangeably by the meteorological community. These are also sometimes referred to as landspouts.
NSTs are considered the most common tornado type worldwide. These tornadoes are typically less violent and occur over a shorter duration than supercell tornadoes (Markowski 2016). However, these tornadoes can easily reach E/F2-rated intensity, as seen in Denver in the 1981 and 1988 cases.
NSTs forming via convergence-vorticity zone processes are common throughout the Great Plains, but have a notable presence in Colorado, Wyoming, and Texas (Bann et al. 2003; Markowski et al. 2016; Wakimoto and Wilson 1989). However, these zones do not produce tornadoes on their own — the presence of a convective storm over the boundary is required, and if you recall, convergence-vorticity zones do not initiate the development of storms without external forcing. Misocyclones along the DCVZ remain as little more than imperceptible wind gyres until a storm or towering cumulus cloud moves or develops overhead. The associated updraft stretches the misocyclone vertically, increasing its rotation speed, which can cause it to develop into a tornado (Wakimoto and Wilson, 1989).
Said differently, once initiation has occurred and individual storms move over the convergence zone, vertical vorticity advection and vertical stretching strengthen the misocyclone to tornadic intensity (65mph or greater).
In general, NST storm reports are generated by observations made from storm spotters or law enforcement. Because NSTs occur in storms without mid-level rotation, they do not trigger automated computer-generated radar-derived Tornado Vortex Signature (TVS) notifications. This results in considerable forecast and warning challenges for meteorologists, especially at DIA.
DCVZ and the Denver Cyclone
There is some ambiguity among prior publications regarding the differences between the Denver Convergence Vorticity Zone (DCVZ) and the Denver Cyclone, both of which typically form over the greater Denver area during the summer months.
The Denver Cyclone is fairly common and is simply a region of weak circulation often forms in the lower elevation terrain of the South Platte River valley to the north of the Palmer Divide. Its presence is often associated with poor air quality in Denver, as stagnant air gets trapped within the cyclone.
With a combination of southeasterly low-level flow (over or around the Palmer Divide) and westerly flow aloft (over the Front Range mountains), the terrain blocking effect between the Front Range and Palmer Divide is what generates this meso-beta scale cyclone over the greater Denver area.
Prior research has found the Denver Cyclone can facilitate the development of a DCVZ-like feature within its broad area of circulation. Wilczak et al., 1991 referred to this phenomenon as the Denver Cyclone Convergence Vorticity Zone. For clarity, we have now mentioned three separate terms:
- The DCVZ, which can form regardless of the presence of a Denver Cyclone.
- The Denver Cyclone, which can form without the presence of the DCVZ.
- The Denver Cyclone Convergence Vorticity Zone (“Enhanced DCVZ”), which is the development of the DCVZ within the Denver Cyclone.
It was later determined that the DCVZ and the Denver Cyclone Convergence Vorticity Zone are indeed the exact same feature. However, the Denver Cyclone’s distinctive flow pattern enhances shear along the boundary, creating favorable dynamics for low level forcing that is not inherently present in non-Denver Cyclone DCVZ cases.
This unique wind field produces a stronger DCVZ (Szoke et al. 2006) capable of enhancing precipitation and even initiating storms along the convergence zone. For clarity purposes, this article will refer to the Denver Cyclone Convergence Vorticity Zone as the Enhanced DCVZ.
According to Walczak & Christian, 1989, the misocyclones along the DCVZ are enhanced by southwesterly surface winds generated by the Denver Cyclone, southerly winds breaching the Palmer Divide, and southeasterly low-level winds originating from the southern plains. Walczak & Christian, 1989, found DMC is initiated by vertical forcing via the southwesterly surface return flow, supporting the conclusions from Szoke et al. (1984): additional forcing is required to initiate DMC along the DCVZ.
Most importantly, Szoke (2006) concludes that during active DCVZ days in the month of June, tornado development at the convergence zone is close to 30%. The chance of tornado development doubles to 60% during a Denver Cyclone. The DCVZ, along with the presence of a Denver Cyclone, are important features for meteorologists to consider when studying and forecasting weather in northeastern Colorado.
Although the Denver Cyclone and accompanying enhanced DCVZ are typically late spring and early summer events, this phenomenon can occur during winter storms. For instance, a snow event on December 4th, 2008 was significantly underforecasted, as it ended up being enhanced by the Denver Cyclone and an associated convergence boundary. The enhanced vertical motion near the boundary amplified snow-liquid ratios (about 1:15 to 1:20 for this event — very fluffy) and increased the amount of precipitation.
More recently, on the night of February 24th, 2021, a Denver Cyclone quickly developed as am expected winter system swept across the High Plains, to the surprise of many local meteorologists (including ourselves). Only a few hours after its development, this Denver Cyclone produced the Enhanced DCVZ which was responsible for snow accumulation rates of well over an inch per hour across parts of the metro area.
In the end, a narrow swath of southern and eastern Denver metro area received an unexpected 12 to 15 inches of snow during what was forecast to be a mild Front Range winter system, with much lower totals immediately outside of the Denver metro area.
Here was the map of totals we posted the morning after the event:
As we enter the spring and summer months, keep your eye out for the DCVZ! Recently, this was visible on satellite on May 19th, 2021.
Wind conditions along the High Plains favored the development of a modest Denver Cyclone accompanied by an Enhanced DCVZ. We were able to grab visible satellite of this event which shows the development of the convergence zone and storms initiating at this boundary. Watch for the explosive development of clouds to the southeast of Denver about halfway through this animation:
Prior events have been difficult to observe on visible satellite imagery due to high clouds blocking the view. This event was particularly clear as weak flow aloft kept mountain convection confined to the Front Range. This Denver Cyclone event initiated a burst of convection along the convergence zone, dissipating completely after approximately 90 minutes. Although no tornadoes occurred with this event, it is otherwise a textbook example of a summertime convergence-vorticity zone.
Written and researched by Laura Smith, with additional contributions and editing by Thomas Horner.
¹ Developed by Dr. Fujita in 1981, the misoscale is a horizontal scale of atmospheric motion which ranges from 4km to 4cm (Murkowski and Richardson 2016). According to this scale, the miso-scale ranges from 4km to 400m, and will be referred to in this article as the misoscale. This scale is smaller than the mesoscale, discussed earlier.
Bann, R., Weiland, M., & McAnelly, R. (2003, April 27). Terrain Influences on Severe Convective Storms Along the Pine Ridge from East Central Wyoming to Northwest Nebraska. NWS Publications. https://www.weather.gov/media/crh/publications/ARP/arp27-03.pdf
Doswell III, C. A. (1980, November 11). Synoptic-Scale Environments Associated with High Plains Severe Thunderstorms. Bulletin of the American Meteorological Society, 61(11), 1388–1400. American Meteorological Society.
Markowski, P., & Richardson, Y. (2011). Mesoscale Meteorology in Midlatitudes. John Wiley & Sons.
Pietrycha, A. E., Manross, K. L., Nelson, E., & NOAA/NWS, Goodland, KS. (2006, November 8). Misocyclone Detection and Observations using the WSR-88D: Operational Implications for the Warning Meteorologist. 23rd Conference on Severe Local Storms. Bulletin of the American Meteorological Society. https://ams.confex.com/ams/23SLS/techprogram/paper_114801.htm
Schreiber-Abshire, W., & Rodi, A. R. (1991, December 01). Mesoscale Convergence Zone Development in Northeastern Colorado under Southwest Flow. Monthly Weather Review, 119(12), 2956–2977. AMS Journals. https://doi.org/10.1175/1520-0493(1991)119<2956:MCZDIN>2.0.CO;2
Smith, L. (2019, May 1). North Eastern Colorado Severe Weather Climatology.
Szoke, E., Cooperative Institute for Research in the Atmosphere (CIRA), Fort Collins, CO, & NOAA Earth System Research Laboratory (ESRL), Boulder, Colorado. (2014, August 21). Landspouts (non-supercell tornadoes) & the Denver Cyclone. DIA Tower Talk. NCAR. https://ncar.ucar.edu/
Szoke*, E. J., NOAA Earth System Research Laboratory/Global Systems Division, Boulder, Colorado, Barjenbruch,, D., Glancy, R., Kleyla, R., & NOAA National Weather Service Forecast Office, Boulder, Colorado. (2006, November 8). The Denver Cyclone and Tornadoes 25 years later: The continued challenge of predicting Nonsupercell Tornadoes. 23rd Conference on Severe Local Storms.
Szoke, E. J., Weisman, M. L., Brown, J. M., Caracena, F., & Schlatter, T. W. (1984, January 14). A Subsynoptic Analysis of the Denver Tornadoes of 3 June 1981. Monthly Weather Review, 112, 790–808.
Wakimoto, R. M., & Wilson, J. W. (1988, November 23). Non-Supercell Tornadoes. Monthly Weather Review, 117.
Walczak, J. M., & Christian, T. W. (1989, November 29). A Case Study of an Orographically Induced Mesoscale Vortex (Denver Cyclone). Monthly Weather Review, 118, 1082–1102.
Wilczak, J. M., Christian, T. W., Wolfe, D. E., Zamora, R. J., & Stankov, B. (1991, August 13). Observations of a Colorado Tornado. Part 1: Mesoscale Environment and Tornadogenesis. Monthly Weather Review, 120, 497–520.