Three major stages of the supercell storm, were observed through ABI high resolution visible and infrared channels.
05 December 2022
28 June 2018
By Ivan Smiljanic (SYSIS)
The western flank of the broad synoptic ridge, which was situated over central USA, was an area of major convective development — an area of sufficient moisture and wind-shear that supported very severe and long-lasting convective episodes. Figure 1 shows the synoptic situation — a high pressure area almost cloud-free with convection development around it. The dotted line in the image shows the boundary of the stable air.
Although the development of the observed system started in the evening of the 28 June, it was interesting to see the two major ingredients for convection, moisture and wind shear, through in satellite imagery earlier that day.
The focus of pre-convective storm assessment is to look for the ingredients in the area of interest. Therefore, the sandwich product with the high resolution visible channel (cloud motion detection) and brightness temperature difference (BTD) of infrared channels (moisture detection) is key (Figure 2).
Note that the focus of this study is not convective system that was already in mature stage earlier that day in the southern parts of North Dakota (development of the system of interest initiates roughly at 20:30 UTC at the west border of North Dakota).
The big difference between two IR ‘window’ channels, namely 12.3µm and 10.4µm, means more moisture in the low troposphere.
The two encircled areas in Figure 2 show the regions of biggest BTD difference, i.e. areas where there is enough ‘fuel’ to support convective processes.
The animation of the same sandwich product (Figure 3) over North Dakota, from the morning hours until the early development of the system, shows slow northward movement of the low level cloud field and ca. westward-moving clouds in the higher troposphere layers. This is the sign of wind shear, both in direction and magnitude.
Confirmation of the low level moisture and vertical wind shear comes from the radiosounding measurement at Bismarck station (roughly in the middle of the state of North Dakota) at 12:00 UTC. Although a small inversion exist at this time, dew point temperatures are relatively high in the lower layers, as well as the CAPE value throughout the upper troposphere.
High moisture content favours the development of the cumulus field in the belt connecting the two regions of higher moisture (Figure 4).
The animation of the visible channel at 500m spatial resolution and 1 min temporal resolution (Figure 5) nicely shows that early development. From the animation it becomes obvious that there are several cloud levels, with the cumulus field slowly advecting towards the north and ‘filling’ the southern part of the area shown.
These cumulus clouds in the south will later become ‘feeder clouds’ for the observed mature system.
Note on the animation that more intense vertical development in the cumulus field is first observed in area 2 (annotated in Figure 2), around 19:30 UTC. Still, these cumulus congestus clouds did not reach maturity, most likely due to a too strong wind shear. It appears that the clouds were ‘blown off’ by the wind and dissipated an hour later, around the time that the main system in area 1 started developing.
Mature convective storm
The final GOES-16 animation (Figure 6) starts less than an hour before initial development of the supercell thunderstorms. Initial clouds associated with this system started to appear around 20:15 UTC and continued to develop up to the higher levels of the troposphere.
Within that hour the storm built to a strong isolated convective cell with an updraft that greatly intensified.
Under existing shear conditions, around an hour after initial development, the intense storm split into two updrafts. These became two storms that were separated relative to the mean wind vector — so called ‘left-mover’ and the ‘right mover’.
A left-moving storm normally contains an anticyclonic mesocyclone with less storm-relative helicity available, being less intense and having a shorter lifetime. That was the case for this storm, where the ‘left-mover’ died off roughly two hours after separation. Separation was also detected by the ground radar system (see animation).
As expected, the right-moving storm further intensified into the supercell storm (confirmed by ground observations) that lasted many hours into the night (for more than seven hours).
Standard cloud-top features, usually observed with such an intensive storm, were detected at a resolution of 500m from the 0.64µm ABI channel: multiple overshooting tops, gravity waves and an AACP (Above Anvil Cirrus Plume) stretching along the upper-level wind field, more than 100km in the wake area of the main updraft.
It is interesting to observe that, while the storm was moving slowly in the south-easterly direction, it consumed the existing cumulus field. As said before, these ‘feeder clouds’ were actually the main driver for the storm (naturally, the ‘fuel’ is the low level moisture that is the reason for the cumulus field's existence).
Note that the black-and-white colour scheme for Figures 5 and 6 is different for visualisation of the mature system (so called ‘Square Root Visible Enhancement’), which is better at pointing out cloud top features. The ‘Linear’ black-and-white scheme was more suitable to show the earlier cumulus field development, providing better contrast between the low clouds and the ground.
As reference for the next generation of Meteosat satellites, with similar imaging capabilities as the ABI instrument, GOES data for this storm was taken as a proxy to envisage the difference between existing MSG and MTG mission capabilities in terms of temporal resolution. ‘Full disc scanning’ and ‘rapid scanning’ for both missions is presented in the 4-panel video in Figure 7.