Official websites use .gov
A
.gov website belongs to an official government
organization in the United States.
Secure .gov websites use HTTPS
A
lock (
) or https:// means you’ve safely connected to
the .gov website. Share sensitive information only on official,
secure websites.
NUCAPS and AllSky both were helpful in tracking the afternoon destabilization over North Dakota today. Both data sources provided what seemed to be fairly accurate assessments of CAPE values. Clear sky and GFS retrievals in AllSky were similar, with the clear sky retrievals very closely matching the NUCAPS retrievals. Having a plan view depiction in AllSky was very helpful when combined with RAP shear vectors in accessing afternoon storm potential.
NUCAPS Sounding:
— warmbias —
ProbTor started giving unrealistically high values in northern ND, noting very high Az Shear Values. This appears to be a problem with data QC in the operational MRMS data, as it does not appear in the Testbed data.
Looking at the time series for one of these anomalous Prob Tor values, its clear the bad Az Shear data is contributing to the jump in Prob Tor:
We continue to see significant issues with ProbTor that seems directly related to inflated Low Level AzShear values. This is likely due to an outflow boundary that is extremely well sampled by the KBIS radar. As this outflow boundary races ahead of convection, the actual tornado threat, which was never high to begin with, continues to drop even more.
ProbTor, on the other hand, places a tremendous amount of weight on LL AzShear, resulting in a persistently high ProbTor values.
#MarfaFront
Watching a supercell near Bismarck…
It was clear from the hail core aloft that this was severe. However, it was obvious around 2152Z that the RFD gusted out as was seen on reflectivity initially, and then later on the velocity images as the winds became less perpendicular to the beam. However, ProbTor continued to carry high probabilities, and even increased when it was clear on the velocity images there was no tornado threat. ProbTor was only keying on the outflow boundary. MRMS Rotation Tracks were also showing a swath of higher values along the outflow boundary.
-Tempest Sooner
Upper left: Orange ProbSevere with a green ring signifying escalated ProbTor
Bottom right: Low Level Rotation tracks with green ProbTor Object associated with the green ring at upper left.
Lower Left: KBIS base velocity.
Key things to notice above – a spurious looking inbound maximum to the northwest of KBIS and a corresponding uptick in Low Level Rotation tracks in response to this shear zone that likely is just an artifact of sampling near the ground. Below, we see ProbTor climbing to over 40% (red trace). Be careful with ProbTor when the storm is close to a radar.
Suggestion: Instead of 0-2km, perhaps Low Level Rotation Track should be capped at the estimated LCL AGL.
#MarfaFront
GLM Flash Extent Density product seems to have poor detection from both GOES 16/17 in the high plains. Perhaps this is related to the greater viewing angle due to latitude and being farther on the periphery. In this example, a very strong supercell was noted to have high total flash density values on ENTLN, while the GLM FED was very unimpressive. It was picking up on some flashes, but not enough to help the storm stand out from a lightning perspective. This could have negative impact on warning forecasters trying to use FED to identify stronger cells from a situational awareness perspective.
— warmbias —
All Sky Layer Cape looking funky with some discontinuities around KS? Why?
A quick look at the Data Type product (below) shows only GFS being used in these suspect areas. Major caveat emptor!
All right, time to look at storms around Bismark.
#MarfaFront
A potential “gotcha” and/or training issue…
I initially loaded a 4-panel with 0.5 Reflectivity and ProbHail overlaid, 0.5 SRM and ProbTor overlaid, and 0.5 Vel and ProbWind overlaid, but I had not loaded the full ProbSevere Model in any of the 4 panels. I then noticed that there was a 29% ProbTor (which was erroneous with bad dealiasing), but my attention wasn’t called to it because the 2nd circle around the storm only appears on the full ProbSevere. It would seem like a good idea to either always have the full ProbSevere displayed or if folks prefer to load the threats separately, perhaps need to add a second circle around the ProbTor object.
-Tempest Sooner
Compared KBIS observed 19z Skew-T to NUCAPS 20z Skew-T
Key points:
Lapse rate differences: Between 1-1.5C/km difference through all levels
SFC and MU CAPE was 500J/kg higher on the NUCAPS
Vertical profile is markedly different and is a much more saturated on the observed Skew-T
To further confuse matters the NUCAPS forecasted SFC CAPE values for the same location and time is 400J/kg…which is closer to the observed sounding than to the NUCAPS sounding.
—Desmond—
Looking north of the ND/MT state up in Canada. Green (Red) is Event Density from GOES-16 (GOES-17) and yellow is ENTLN cloud to ground strikes. Remarkably good agreement in position, although the data smearing is evident and parallax isn’t much of an issue (yet?) given the shallow convection.
#MarfaFront
