ProbSevere and Lightning Jump - Great Match for Warning Confidence

Another cell developing to the east of the two previously warned cells provided another opportunity to use both ProbSevere and Lightning Jump Detection Algorithm in conjunction to issue a Severe Thunderstorm Warning.

Really started to see an increase in the mid level reflectivity on the storm and during that time the ProbSevere increased from 8% at 2204Z to 64% at 2208Z.

At that point, we were starting to think about issuing a SVR and as we were making that decision, there was a 5 sigma lightning jump noted. That data arrived before the next version of ProbSevere arrived but tied in with the increase shown in the previous couple scans. At that point, we decided to issue the SVR and our thought was confirmed by the next version of ProbSevere increasing to 84% with a MESH value of 1.25".

The warning was issued at 2217Z and the ProbSevere value reached 94% with a MESH of 2" as the warning was issued.

Although the experience with these two products has been limited thus far, they have shown to work well in conjunction with the radar analysis to lead to increasing confidence on issuing a warning during rapid intensification.

-SRF

KICT Experimental Severe T-Storm Warning

The Omaha team moved to Wichita, KS WFO and has focused on developing convection along a cold front.  The NOAA/CIMSS ProbSevere model, combined with analysis of base radar data, helped the forecasters issue an experimental severe t-storm warning at 2217 UTC 04 May 2005 on the cell centered in Figure 1.  The animation shows how the ProbSevere model probabilities rapidly increase as the observations of the storm intensify.  The forecasters commented they liked how the rapid uptick in ProbSevere output matched with their analysis of the 3-d reflectivity of the storm as well as 5 sigma lightning jump (not shown) increased confidence to issue a warning.

Figure 1.  NOAA/CIMSS ProbSevere and MRMS Composite Reflectivity valid 2200 - 2214 UTC 04 May 2015.

-Sieglaff

First Warning for the Project (ICT)

With the move to ICT we were able to issue our first warning of the project. The storm that was previously discussed re-intensified after splitting. The increasing reflectivity aloft led to the MEHS data to give a hail estimate of 1.55" and pushed the ProbSevere up from 58% to 95%.

Thus, a warning was issued for that cell.

To the northwest, a storm in DDC started to take a right turn right around the same time as a series of 2-3 sigma lightning jump values (orange in bottom left screen). Those values occurred as the storm intensified on radar and led to the ProbSevere values rising from 80% to 97% as it was nearing the cwa. Also, notice the overshooting top detection in the lower right screen.

With all of those values and the expected movement into the cwa, issued a warning for the ICT area for the storm.

-SRF

Convective Initiation Success!

Looking back at the big picture and focusing on the GOES-R convective initiation, there was not much showing in our area of responsibility (TOP) but further south in northern Oklahoma I noticed some increasing probabilities at 1900 UTC.

15 minutes later at 1915 UTC the CI probability jumped to 90% on the east side of a cluster of identified clouds while there were also 40% probs on the west side.

Immediately after this time it was evident there was convective development was occurring as seen in this 2000 UTC satellite image below.

By 2030 UTC some CG lightning strikes were noted seen in the image below.  Radar at 2030 UTC also included.

Jack Bauer

Moved to ICT

With limited convection in the OAX area, we have moved over to ICT. The ICT CWA is along the same boundary, but now is on the southern side.

Taking at the quick look at the radar, there is a strong thunderstorm over the north-western part of the CWA. I decided to use the ProbSevere data as a quick overview and found the MRMS MESH data (hail size ~1") to be very helpful in giving a quick look at the current expectations on the storm. I also found the satellite indicators to be helpful with the switch to indicate the current state of the storm (Mature).

Starting to see some increase in the eastern cell while typing this and will follow up this blog post with additional information. -SRF

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.