Exploring the Meteorological Environment of the Longview Tornadic Supercell of July 2, 2016

This post is an exploration of the synoptic and mesoscale background environment surrounding the Longview tornadic supercell of July 2, 2016, from which we observed two tornadoes – both given a final rating of EF0. The tornadoes of July 2 were just two of five confirmed supercell tornadoes in Alberta over 4 consecutive days (June 30-July 3). A southwesterly jet axis provided the deep shear necessary for organized convection, which overspread rich low level moisture throughout the period. Several well-timed disturbances tracked across Alberta, aiding in the development of daily severe convection near or just after peak heating.

The synoptic environment would become supportive of supercells on the afternoon of July 2, owing to the interaction of a shortwave trough with a moist, moderately unstable air mass to the east of the Rockies in southern Alberta (Fig.1). MLCAPE values were forecasted to be between 1000-1500J/kg (Fig. 2), and 0-6km bulk shear would become about 30-40 knots as the speed max associated with the disturbance tracked over the area (Fig. 3). However, there were no pressing indications that tornadoes were likely on the large scale, due to the absence of any major synoptic features that could enhance low level shear. These would have to be sought out on the mesoscale – and indeed, it did appear that some smaller scale features played a direct role in the development of a discrete, lone tornadic supercell that was likely a “mesoscale accident”. As such, no tornado watches were (understandably) in place at the time the first tornado was observed. My personal theory of what the mesoscale features may have been that ultimately contributed to tornadogenesis will follow.

Looking back, I wish I had undertaken this account immediately after the event, so I could have had access to better mesoanalysis data, visible satellite imagery, synoptic charts, and screen grabs of close-up radar scans. As such, I have been limited to using low resolution mesoanalysis data from the SPC meso archive, low resolution historical radar scans from ECCC’s public climate website, historical hourly obs from sites in the storm’s proximity, a personal forecast discussion written that day (which mentions the possibility of tornadic storms with boundary interactions – Fig. 4), and personal photographs from our documentation of the storm.

On the evening of July 1, a relatively slow-moving disturbance contributed to the development of a rather large cluster of convection that would track across southcentral Alberta well into the overnight hours. Early in the event, a confirmed tornado (EF0) occurred near Bergen in the mid-evening. The convective cluster would kick off a substantial cold pool that tracked well to the south of Calgary in the overnight hours, before likely coming to rest along the foothills, and extending southeastward over the southern Alberta plain. The cool air mass in the wake of the outflow boundary would begin to modify with daytime heating on July 2, but an enhanced area of moisture convergence would occur along the residual boundary, resulting in large part from a baroclinically-driven circulation that is characteristic of such mesoscale boundaries.

A well-defined outflow boundary is visible tracking south, to the south and east of Calgary at 3am. This likely pushes further south before coming to rest along the foothills near Longview, and extending SE from there.

A relatively narrow corridor of high boundary layer relative humidity developed along the cool side of the boundary, where several stations reported dewpoints in the mid-teens along with temperatures in the low 20s – yielding temperature-dewpoint depressions in the 6-8C range. A ways north of the boundary, temperature-dewpoint depressions were frequently near 10-12C within the modifying cold pool, and up to 15C south of the OFB. A quasi-stationary dryline also existed near the front range of the Rockies, evidenced by dewpoints in the low single digits at Bow Valley in light west-southwesterly flow, compared with low teen dewpoints at Springbank airport roughly 40km to the east, which was in moderate east-southeasterly flow.

My re-analysis of surface winds, temperature, and dewpoint at 21Z. Approximate location of dryline (brown) and outflow boundary (blue); red dot marks the location of the strongest radar echo at 21Z, and the two green dots mark the location of the confirmed tornadoes.

Convection initiation first occurred near 1900 (times hereafter documented in UTC), with the first radar echo being detected by XSM (Strathmore radar) at 1920, about 20km west of Millarville. The storm would go on to propagate very slowly to the south-southeast for over two hours, as deep layer shear was still relatively weak some distance ahead of the approaching shortwave trough. During this time, a weak mid-level mesocyclone was observed on radar, and we visually observed some marginal supercell characteristics. However, the storm appeared to be struggling at times, with varying degrees of precipitation intensity, and periods of time where the storm appeared to become quite linear. One thing that was striking during this time however was the strength of the sustained wind in the storm’s inflow, as well as how low the LCL heights were due to the rich boundary layer relative humidity. While there were a few observation sites upstream of the storm, the overall low density of sites in southwestern Alberta would render the finer details of the wind and moisture shrouded in mystery.

The initial cell fired early in the afternoon, and quickly reconstituted itself to the ENE by 2000. It then moved very slowly to the south over the next two hours, before suddenly intensifying, with an attendant increase in storm motion. Once the storm became elevated, it was advected eastward with the mean cloud layer wind.

I wish I'd had a way of measuring the inflow wind speed, but I estimate about 30km/h sustained here. 2230

Low level clouds are seen condensing along the low level inflow current

Rather suddenly, the storm ramped up in intensity near about 2230, only about 30km southeast of the position of its origin. The speed max associated with the disturbance likely arrived near this time, increasing the deep layer vertical shear, which subsequently led to the strengthening of the mid-level mesocyclone. However, given that the 0-6km shear vector was approximately west-southwesterly, we were surprised by how hard of a right angle the storm motion vector was (about 135 degrees/southeast). This, combined with a sudden increase in intensity, leads me to believe the storm’s propagation was being governed by more than its own dynamics; the position of the pre-existing NW-SE oriented stale OFB was likely also playing a pivotal role. As the storm propagated down this boundary, it would readily ingest a source of streamwise vorticity roughly aligned with the storm relative inflow, that would greatly increase the strength of low level mesocyclogenesis.

At about 2250, we observed a dry slot developing, associated with the first occlusion downdraft. Very shortly after, at 2255, we reported the first funnel from about 2 kilometres to the southeast. (It turns out that this funnel was associated with a ground circulation, evidenced by downed trees we found in the area after the fact). The tornado was short-lived, but a ground-scraping wall cloud remained. Then, about 15 minutes later, a large funnel began to hang over the fields below as a part of a new cycle. This funnel became rain-wrapped as well, but we were able to see it in contact with the ground for a brief period of time. The storm’s constant tendency for having wet RFDs (evidenced by a fat hook echo seen at the time of the second tornado on a Radarscope screen capture) likely stemmed from the relatively weak anvil-level storm relative winds (Fig. 5), which were insufficient in venting precipitation further downshear of the main updraft. After losing our visual and being hit by the storm’s warm RFD winds, we raced east of Pekisko to get back into a favourable viewing position in the storm’s warm sector.

Tornado #1, 2255

Trees down just east of Highway 22, south of Longview, where the first funnel was sighted.

At times the wall cloud was nearly scraping the ground. Here, a large funnel is very low to the ground (2316), with the most visible mid-level rotation I've ever seen.

A rather large funnel characterizes tornado #2, at 2318

Base reflectivity scan at 2320, during tornado #2. Pronounced V-notch and fat hook echo apparent.

When we arrived at our new position west of Cayley just after 2330, the storm suddenly began to lose its structure, and appeared to rapidly become elevated. We surmise that this is because it moved off of the enhanced convergence zone near the OFB/foothills, and into a strongly capped environment over the plains to the east (already sitting in a cool, more stable modifying cold pool). The storm nearly vanished from radar, before picking up somewhat and being maintained thereafter as an elevated thunderstorm associated with the disturbance tracking across southern Alberta into the late evening hours.

In summary, from a nowcasting or chasing standpoint, it is worthwhile being aware of the subtleties that may exist on the mesoscale in greater environments favouring the development of deep, moist convection. In our case, the previous night’s convection would play a role in the following day’s setup, since it would lay down an OFB that would work constructively with an existing thunderstorm, as well as prevent convection further within the stable air of the modifying cold pool. Stale OFBs are a source of low level convergence, moisture, and shear (stronger low level shear results over the cool side of the boundary where the boundary layer depth is shallower), and knowing its position can help us anticipate storm behaviour in its vicinity. In our case, the initial storm may have developed near the intersection of the OFB and dryline – though afterward, it appears the dryline played no further role in influencing the storm. The cool side of the OFB also provided an environment conducive to tornadogenesis, since the necessary ingredients of stronger low level shear and rich boundary layer relative humidity were present there. The localized small temperature-dewpoint depressions allowed for a much lower based storm than all others further north along the foothills that day, which provided a clue that something special was going on there.

Figure 1: 500mb heights at 21Z revealing a shortwave trough to the immediate west

Figure 2: MLCAPE at 21Z revealing moderate instability over southern Alberta. It was very much capped, despite the apparent lack of MLCIN on this image.

Figure 3: 0-6km Bulk Shear at 22Z. Marginal for supercells, but doable, as it turns out.

Figure 4: My CFD for July 2.

Figure 5: Storm relative anvil level winds were quite weak, at under 20 knots.