Heavy precipitation and flooding are not uncommon for the state of Colorado. The combination of moisture from the seasonal monsoon and upward lift by the complex regional topography can trigger convection and provide the necessary ingredients for heavy rain-producing storms.
On 9–16 September 2013 large-scale weather patterns led unseasonal amounts of moisture to be transported from the Gulf of Mexico and eastern Pacific Ocean toward the Colorado Front Range, resulting in historic flooding throughout the state. On the evening of 11 September and into the next day, Boulder, Colorado experienced extremely heavy rain rates, which resulted in more than 180 mm of rain accumulating in a six-hour period.
At the same time, local radar identified a cyclonic circulation, or mesoscale vortex, that originated near Denver and traveled northwest towards Boulder. Strong low-level easterly (east to west) winds north of the vortex suggested enhanced upslope flow, resulting in enhanced convection and precipitation. Those observations suggest a plausible connection between the mesoscale vortex and flash flooding in Boulder and surrounding areas. This led my colleagues and me to wonder: How did that vortex form?
Figure 1. Radar observations from the Denver (KFTG) radar at 0612 UTC 12 September 213 showing the mesoscale vortex over Boulder. (a) Radial velocity showing a couplet of winds toward the radar (green) and away from the radar (red) in close proximity. The couplet is associated with rotation. The blue arrow represents the direction of upslope flow. (b) Radar reflectivity showing an enhanced convective band over Boulder. The cities of Fort Collins (F), Boulder (B), and Denver (D) are provided as reference.
The mesoscale vortex originated in a region associated with a lee vortex phenomenon called the Denver Cyclone. The link is plausible, but due to the large amounts of phase changes (that is, water vapor being converted to liquid water) that occurred. However, we hypothesized that latent heating played a more important role in the development of the September 2013 vortex than did lee vortex formation.
Latent heating can influence the generation of vortices through conservation of potential vorticity (PV). Between two isentropic (constant potential temperature) surfaces, PV is conserved. Latent heating causes buoyancy differences that lead to upward motions. Those motions, in turn, transport mass across isentropic surfaces, resulting in less mass and more PV below the source of diabatic heating, and more mass and less PV above the source. In that way, a positive PV anomaly and associated cyclonic circulation will be induced below the convection. This mechanism has been identified in past studies of mesoscale convective vortices (MCVs), which often occur in the Great Plains of the US, as well as other parts of the world.
Numerical simulations
Atmospheric scientists use numerical models to explore the physical processes behind meteorological phenomena. We did the same to investigate the mesoscale vortex observed over Boulder on 12 September, and to test our hypothesis that latent heating was behind its formation.
Our numerical simulations were performed using the Advanced Research-Weather Research and Forecasting (ARW) model. The first experiment turned latent heating off for the entire simulation (LH_OFF) to determine if the release of latent heat had any effect on the development of the vortex. The results of that experiment showed no vortex development and lacked strong low-level easterly winds. The findings suggest that latent heating did play a large role not only in the development of the mesoscale vortex, but also in the low-level easterly flow, which was associated with upslope flow and enhanced convection.
The Denver Cyclone typically forms during southeasterly flow, which was observed during both the control and LH_OFF simulations. Given that the LH_OFF run did not reproduce the September 2013 vortex, we concluded that that lee vortex formation associated with the Denver Cyclone was not the primary mechanism during that event.
Figure 2. Comparison of the control simulation and the experiment where latent heating is turned off for the entire simulation (LH_OFF). Maps of potential vorticity (colors filled in PV units), wind barbs (knots), and easterly wind contours (black contours starting at 10 m/s at 5 m/s intervals) at 1-km above ground level (AGL) are shown for (a) the control simulation and (b) the LH_OFF experiment. (c) Shows the easterly wind difference between the control and LH_OFF at 1-km AGL (m/s), with red (blue) colors depicting a net easterly (net westerly) difference.
Two other experiments were performed in which latent heating was turned off either for the first 18 hours of the simulation or for the last 42 hours of the simulation. The two experiments aimed to determine when latent heating from precipitation was most important in the development of the September 2013 vortex. The results from these simulations suggested that the latent heating just before and during the event was more important.
The three experiments led us to conclude that latent heating was indeed a critical mechanism in the development of the mesoscale vortex and influenced the low-level flow that, in turn, impacted convection and precipitation.
In lieu of high-resolution observations, our numerical study used the control simulation to explore the distribution of latent heating within the vortex’s environment. Latent heating profiles show that on 12 September (when the vortex was observed in the model) the atmosphere had convective characteristics—that is, a strong, positive vertical gradient in latent heating in the lower troposphere. A strong gradient leads to the development of a positive potential vorticity anomaly, which is associated with a cyclonic circulation.
Further data analysis found that the conversion of water vapor to liquid water droplets (cloud water condensation) was the dominant microphysical process in the lower troposphere. Higher up, the deposition of water vapor to snow predominated. The large amounts of cloud water condensation resulted in the positive vertical gradient observed in the control simulation on 12 September.
Figure 3. Time–height plot of accumulated latent heating (color filled; kelvin) and potential vorticity (black contours starting at 1 PVU at 0.5 PVU intervals) for each hour in the control simulation. Note, the vertical axis is geopotential heights in km mean sea level and is stretched to emphasize the midlevels of the troposphere. 12 September encompasses forecast hours 24–47.
To test the sensitivity of the mesoscale vortex’s development to the strength of cloud water condensation, we performed a final experiment in which we reduced the contribution of that dominant microphysical process by 50%. The experiment resulted in a huge reduction of the vertical gradient in latent heating, which produced neither a vortex nor enhanced low-level upslope flow. The precipitation over Boulder and the northern Colorado Front Range was cut by 63%.
Taken together, the results show that a positive feedback mechanism was in play during this event: Cloud water condensation led to the development of a positive vertical gradient in latent heating, inducing the cyclonic circulation and low-level easterly flow, which resulted in enhanced upslope flow, triggering convection and more cloud water condensation.
The ARW simulations and sensitivity studies presented here suggest that the mesoscale vortex was indeed responsible for the increased rain rates observed over Boulder on 12 September 2013 that led to the severe flash flooding. The mechanism involved in the vortex’s development was more akin to a mesoscale convective vortex than to the Denver Cyclone.
The results also show the importance of properly representing microphysical processes and their contribution to latent heating within a numerical model. If the model misrepresents them, their contribution to latent heating will be altered and the development of circulations, such as the one studied here, could be missed in the forecast, leading to an underestimate of the precipitation. A thorough understanding of the processes that led to the development of this mesoscale vortex will help atmospheric scientists better forecast these phenomena, their subsequent effects on precipitation, and their impact on public safety.
Annareli Morales is a PhD student in the Department of Atmospheric, Oceanic, and Space Sciences at the University of Michigan in Ann Arbor. This article is based on work performed for the author’s MS thesis at Colorado State University, with the support of Russ Schumacher and Sonia Kreidenweis. A manuscript of this work has been submitted for peer review to Monthly Weather Review.
