CI before severe thunderstorm

CI was 81 at 1830z in northern Jeff Davis County (first image). By 1901z 0.5 deg reflectivity (image 2) showed the first greater than 35 dBZ returns. By 1945z the Probability of Severe was 27 percent (image 3). Severe hail, 1.75 inches, was reported at 2008z through 2025z (image 4). By 2049z the storm had a massive Three-Body Scatter spike (image 5).

Time from the 81 percent chance of CI to hail report was 98 minutes.

Time from the first 35 dBZ returns to hail report was 67 minutes.

Time from 27 percent Probability of Severe to hail report was 23 minutes.

-Lynford

Last three images from Vivian Darkbloom

Prob Severe Color Curve/Polygon thickness

I like the gradual color change and polygon thickness increase as the probabilities increase. See the image below for an example of the lower prob (thin blue) vs the higher prob (thick pink) and some gradients in between. This allows a warning forecaster the opportunity to quickly filter the sub-severe storms and focus on the potential for stronger convection in other areas.

Brick Tamland

Storms over Midland TX WFO

Storms quickly intensified quickly Monday afternoon over the Midland, TX WFO.  The first storm to exhibit high NOAA/CIMSS ProbSevere values was west of Fort Stockton, TX (83% 1950 UTC 04 May 2015, Figure 1).  This storm continued to intensify and was warned by the Midland WFO at 2016 UTC 04 May 2015 (Figure 2).  The storm produced golf ball sized hail at 2020 UTC.

Figure 1. NOAA/CIMSS ProbSevere and composite reflectivity valid 1950 UTC 04 May 2015.

Figure 2.  NOAA/CIMSS ProbSevere, NWS Midland Severe Thunderstorm warning and composite reflectivity valid 2016 UTC 04 May 2015.

Of note with this storm was the presence of strong satellite-based glaciation rates, which helped provide the 26-minute lead-time prior to the first warning and 30-minutes to the first severe hail report.  The EWP forecasters noted the good lead-time of the ProbSevere product for this storm.

-Sieglaff

A Quick First Look at GOES-R CI

Forecasters are working in the Midland, Texas county warning area investigating all of the products that are available to them for evaluation this week.  Like my earlier post on lightning observations available at HWT, this post is a quick highlight of the GOES-R CI.

The image below came from 4 May 2015 at 2000 UTC.  The GOES-R CI (upper left) highlights a region with 60+% probabilities.  The upper right shows the Earth Networks 1 minute lightning observations while the lower left shows the radar reflectivity.

GOES-R CI at 2000 UTC for 4 May 2015 in the Midland, TX CWA.

By 2036 UTC, a small cell with 40+ dBZ is observed (circled) and shown below.

GOES-R CI at 4 May 2015 at 2036 UTC showing initiation 36 minutes after CI indicated 60% probabilities.

ProbSvr and the Development of a Supercell in Northeastern Kansas

The storm of interest began to develop in a weak line of showers around 1930 UTC.  Surface obs through the area indicate temperatures in the lower 80s and dewpoints in the upper 50s to low 60s.  ProbSvr indicated around a 10% chance of the storm becoming severe.  The storm slowly strengthened over the next 20 minutes, but the Prov Svr remained around 10%.  From 19:52 to 19:57 UTC, the storm intensified rapidly, and the Prob Svr jumped from 13% to 72% (time of two radar scans)

It appears the storm is beginning to root along the E-W warm frontal boundary through the region.  At 20:11Z, ProSvr MRMS MESH indicates 2" in diameter hail in the highest dBz core.  This area correlates well with lower values of Zdr of near or just below 0 dB.

Although I did not see any hail reports while I was watching the storm, the first severe wind gust report (60 mph) occurred at 20:30 UTC, about 32 minutes after the ProbSvr rapidly increased.  Because I used the site radar (KTWX), there was a slight lag in the availability of the ProbSvr data.

Severe Storm in Midland's CWA

Now that we have a verified severe storm in Midland's area, let's see how the lightning jump and probability severe algorithm worked.

At 1945Z, the algorithm showed a severe probability of 27%. The lightning jump was nothing to jump about.

With the next 0.5 deg scan at 1950Z, the probability skyrocketed to 83%.

At 2031Z, a 3 sigma lightning jump occurred. Severe probability was 93% at this time.

By 2049Z, the cell grew into this beast. Max reflectivity on the 0.5 deg tilt was 75.5 dBZ @ 16000 ft MSL/12230 ft AGL. A prominent TBSS had developed on the storm by 2040Z.

The LSR associated with this storm is shown here. Severe hail was observed beginning at 308 PM CDT/2008 GMT. In this case, the lightning jump occurred over 20 minutes after the storm began to produce severe hail. However, the probability severe did provide a lead time of about 20 minutes.

-V. Darkbloom

NUCAPS Sounding for NE Kansas

With the main convection still to the south of the area, decided to test out the 1900Z NUCAPS sounding. With clouds and convection closer to the KS/NE border, went a little farther south into Kansas for the sounding, which conveniently lined up the KFOE site.

Clicking on that point came up with the sounding shown below.

This sounding provided a good first guess, but looking at the surface data (75/56 on the sounding) indicated that it needed to be adjusted to match the ob at KFOE (84/59).

Making that adjustment to the surface conditions increased the CAPE values from around 1350 J/kg to 2660 J/kg. That put the values better in line with the initial values see on the RAP initial conditions. -SRF

Lightning Observations at HWT

Total lightning has a large role in the Experimental Warning Program this year.  We are on day 1 of the first week and everyone is familiarizing themselves with the various products.  I have shown an quick comparison of some of the products that we will be evaluating throughout the week.

The storms cooperated by moving into the West Texas LMA domain (Lubbock county warning area) allowing for the first views of the NASA SPoRT pseud-geostationary lightning mapper (PGLM) this year.  The image shows the PGLM (upper left), radar reflectivity (upper right), the lightning jump showing the sigma level change using Earth Networks data (lower left), and the Earth Networks total lightning over a 5 min summation and put on an 8 km grid.

The PGLM is showing a maximum value of 41 flashes over a 2 min window with the Earth Networks showing ~40 flashes over a 5 min window.  The lightning jump indicated the storm in the north as having a 1-sigma change, which is not enough to indicate severe weather is imminent.  The PGLM gives another view of spatial extent of the lightning, which is particularly evident in the northernmost section.  The main advantage of Earth Networks is the coverage, unlike the small domains from the lightning mapping arrays that are used to derive the PGLM.

Lightning observations from the 2 min PGLM (upper left), radar reflectivity (upper right), lightning jump algorithm (lower left), and Earth Networks 5 min data on an 8 km grid.

First Severe TOP

First severe storm warning issued at 2002 UTC for a storm over northern Pottawatomie county along strong instability gradient, just southeast of a weak surface low and in the vicinity of weak surface boundary.

Looking at the all-tilts products we noticed 60 dbz values above the -20C level and the ETN displayed a DTA with a significant increase in flash rate from around 30 to 60 to over 100 flashes in consecutive scans.  This coincided with GOES-R overshooting top detection as well as the lightning jump algorithm showing a 2 sigma jump followed by a 6 sigma jump.

The probability of severe increased to over 90% shortly after the warning went out and the PHI also was showing values in the 70s preceding the warning and was above 80% at the warning issuance.  This particular case in my mind really builds my confidence in the potential of the lightning jump, PHI and prob severe algorithms as it relates to warning operations and potentially improving lead time, of course after proper studies have been done in both verifying and null events.  It would be helpful to have some sort of threshold values for PHI and Prob severe as general guidance (based on geographical location).

DTA track was not very representative of the actual storm motion as it had a northern bias.  This could be due to the storm favoring the instability gradient and local boundary as the actual storm motion was more east of the DTA track.

Jack Bauer/Brick Tamland

AMA is Lighting Up

The excitement at the moment is in the Texas Panhandle. A quick look at one of the cells in Amarillo's CWA:

(note: this is LBB radar data as we're not pulling in AMA data at the moment)

The ENI Lightning Jump algorithm shows a 2 sigma jump. The contours are the NOAA/CIMSS highly experimental probability severe, coming it at 47%.

It had peaked slightly earlier around 60%.

19Z temp and moisture profiles from JPSS came in just after 20Z. I picked a point sounding in the relatively clear area of the eastern OK Panhandle...

And the resulting temp/dewpoint profile:

Surface based CAPE was 1016 J/kg with an LI of -4C. This is comparable to the approx 1100 J/kg and -5C shown on the GOES-R LAP algorithm in this area at 19Z.

WFO AMA ended up issuing a severe thunderstorm warning for this storm. Now we wait for verification.

-V. Darkbloom