Wed forecast for ICT first thoughts

Developing storms to the west expected to continue to grow in coverage and strength as they move NE into ICT area this afternoon. Instability remains very high w/ MLCape ~2500 and deep layer shear in excess of 50kt. Soundings from LMN and OUN at 17 & 18Z showed a small cap further south with better low level turning. Soundings further north were more straight and slightly longer indicating splitting cell potential at ICT. A weakness in winds around 700 mb remains a concern through the day as the LLJ strengthens near sunset and enlarges the lower part of the hodograph. Discrete storms will be hard to maintain but there could be one or two in advance of storm clusters moving east late in the afternoon. Main threat will be hail and wind with the best tornado threat across the western 1/3rd of ICT's area before mode goes linear. - S Coulomb

Week 1 Day 3 Operations

Looks like we are in for another busy day today operating in the Norman, OK and Wichita, KS CWAs

Week 1 Day 2 Wrap Up

The forecasters finished up a long busy day yesterday with numerous storms occurring across central Kansas most of the day. Feedback from the day is posted below.

• ProbSevere
  Seemed to pick up on severe weather earlier than the radar and MESH seemed to pick up on the severity of the storm
  ProbTor picked up on a small tornado near Buffalo, Ok where it increased to 70% in just a few minutes
  Good for situational awareness on the busy day that was yesterday with a lot of storms around, it helped pick out storms to examine more
  Probabilities seem to be really well calibrated to the severe probabilities
  ProbSevere stayed high on hail on a left split supercell even though MESH dropped off some
  Probabilities picked up on the increased lightning before MESH picked up
  Still working on optimizing the display to be able to view ProbSevere and ProbTor displayed together
  Gives some increased confidence to warn a little earlier than might would have with just radar
  
  
  CI
  Seemed to really have problems along the dry line
  Stayed low the entire time
  Had a problem in the code that is looking to get fixed ASAP
  
  
  GLM
  Flash rates seemed to be lower than expected compared to some of the other networks
  Didn’t notice any nice smooth trends up or down than what was expected
  A lot of the data use was really a learning experience using some of the new data sets
  Values in the total energy did seem to spike up prior to severe reports and tornadoes
  Some of the color table cut offs could be re done for their to be more contrast
  Would like to either have some transparency or more ideally a contour to be able to overlay on satellite data and still see the storm tops
  Being able to combine everything into one panel would be ideal to save screen real estate when in operations
  Quite a bit of training will be needed to get forecasters up to speed and to understand what they are viewing when they use it operations
  Training on uses in different environments or storm modes would be ideal to train up forecasters on the uses
  What definitely needs to be trained on is the benefits of what GLM gives you that other ground based network doesn’t
  Having job sheets for live weather of what to look for in the lightning data while using it
  Need to build off of applications of the data
  Integrating with other lightning data is important
  
  
  NUCAPS
  Modified wasn’t making modifications to some of the profiles that needed to be modified
  The surface data should still at least be modified based on the RTMA data even if nothing else is
  Cloudiness yesterday caused some trouble with most of the soundings
  Gridded data wasn’t really used at all with all of the cloudy areas around yesterday
  Forecasters had to modify themselves since the modification wasn’t working on some of the soundings
  The profiles seemed to match pretty well in the mid and upper levels to the 19Z special sounding from DDC
  
  
  ABI and All-Sky LAP
  Baseline stability indices were blocked a lot yesterday, but the all-sky LAP seemed to match up pretty well to the CAPE expected over the area
  The satellite retrievals seemed to bring down the GFS first guess too low by about 500 J in most places
  Split Window difference showed some higher values ahead of and behind the dry line
  The derived winds compared to the special sounding were slightly weak from 500mb down to the surface
  Could make out the dry line boundary in the split window difference with low level moisture pooling on the boundary
  Looked at layered precipitable water to look at increased moisture above the boundary layer and the mid level moisture started increasing really fast in central Kansas. The convection developed right along the moisture boundary
  Cloud Phase RGB really picked up on the initiation but it could also be picked up on in the visible imagery
• 
  

Michael

Storm split represented in GLM

As the supercell moved out of OK it shed a left split before turning more east. The GLM captured this split with higher event density values associated with each updraft core. The GLM flash event cores became increasingly separated as the left mover moved north. - S Coulomb

Active Severe Weather Day in Hastings, NE

ProbSevere was very helpful for hail and wind, but especially hail, which is the main threat as of late afternoon. ProbTor was lagging behind base data by a volume scan, which was critical as a tor warning was issued several minutes before probtor spiked up from 7% to 46%

GLM 1-min Total Energy overlaid with 1-min vis cued me in on areas of deep convective development. Additionally, ENI lightning data overlaid with GLM gave me a good indication of where anvil flashes were propagating away from the main updraft.

GLM readily caught my attention as energy was displayed as an image, with more intense lightning activity showing up 'hotter' in the magma color table.

Supercell that moved eastward just north of the KS/NE border exhibited high total energy while located in our CWA. It appeared that as the storm's total energy began to weaken as it moved away from our CWA, a confirmed tornado was reported in Gage County NE at 7:23PM local time (see bottom two images). Lightning jump?

First look at multiple GLM products!

Clockwise from upper-left:  GLM Total Energy, Flash Extent Density, and Flash Area.  ENI is overlaid with Total Energy and Flash Area.

Saw an interesting spike with the cell near the center of each image, a couple counties south of Hastings CWA outlined in pink.  The spike with this cell happened near the end of the loop after 2200Z.

My warning partner Tom announced that the GLM experimental one minute data was available while I was looking at some 5-min visible.  Excitedly, I overlaid the one minute lightning on top of the satellite (time matched) not realizing I might see something tantalizing, so only got 5 minute updates of the one-minute lightning data.

A strengthening storm just south of Hastings CWA showed a sudden spike in Total Energy (TE) at 2217Z.    A sequence from 2212-22Z saw TE bounce from 34 - 111.7 - 40.5fJ.  Flash density remained 5-10 throughout.

Coincidentally, there was a spotter report of hail 2.5" at 2212Z with the cell.

I enjoyed the discussion I had with the developers today, and appreciate and like the new color scales they quickly developed for us since I had made this above gif.  Tomorrow, I will revisit GLM, hopefully with better planning!

-Forrest

GLM Fields for Severe Thunderstorm

Mature storms northeast of our forecast area later in the day provided a great demonstration of a few of the lightning fields and how information from them is relevant. The Flash Extent Density and Total Energy (top left and bottom left) highlight the regions of greatest lightning activity corresponding to strongest and most persistent updraft regions. Further, Average Flash Area (upper right) further delineates updraft regions by highlighting the smaller flash areas. Brief lightning activity is sometimes apparent out in the anvil regions, and have appropriately larger flash areas.

-Bucky

Convective Initiation in VC of Hastings (Derived Products and LPW, and mention CI)

Convection rapidly initiated along a frontal and sharp moisture boundary in the Hastings CWA near 2000Z.

In the animation below, LPW 0.9-0.7 rapidly increases, from near 0.39 in the area near 1630Z, to as high as 0.6 by 2000Z, then quickly to 0.65-0.71 only 30 minutes later.  Initiation was very near 2000Z.

Before 2030Z, the area of maximum LPW had relatively thin cirrus moving over.  After 2030Z, thick anvil quickly moved over the area of maximum depicted moisture.

ENI lightning is overlaid.  Lightning first appears at 1958Z.

On a side note, I initially launched LPW stand-alone.  The loop had some blank frames.  On closer inspection, updates happened usually every approx 30 mins, but occasionally, a frame would update only one minute after the last, causing the blank frame.

The increasing moisture aloft was paralleled by increasing convective parameters as seen in the derived products.

K-Index, a parameter sensitive to mid-level moisture rose from the low-mid 20s before 16Z, to a more convective friendly 30+ by near 1830Z, to remarkably high values near to above 40 by 2000-2030Z.  Changes to derived CAPE and LI were much more modest.  In the 4-panel below, clockwise from top-left is:  derived CAPE, Layer Cloud Type, LI, K-Index(bottom left).

The K-index was also diagnosed as increasing quickly on RAP, but overall values in the model were lower.  As seen on the SPC RAP MesoAnalysis, the initial development happened along the moisture gradient at 925mb.

This lends hope that these products will be useful over areas that RAP analysis does not cover, such as oceanic areas.

As a last note, CI didn't show a strong signal beforehand in the area of initiation.  There were some transient values above what seemed to be widespread unfocused nominal values 0-20%, but it didn't clearly key in on the cells that went on to develop rapidly.  As mentioned earlier, there was some cirrus in the area leading up to development, but it did not seem very opaque.

-Forrest

Spin up of another tornado from a long-lived supercell

A long-lived supercell thunderstorm transversed across the northern Wichita, KS WFO and into the Topeka, KS WFO Tuesday evening.  This supercell had exhibited ProbSevere All Hazards probability of tornado values that cycled with the intensity of the low and mid level circulation.  Between 2342 UTC and 2358 UTC the probability of tornado increased from near 50% to over 90% as both low and mid-level rotation quickly strengthened (Figure 1).  This storm produced at tornado at 0004 UTC 02 May 2018.  Forecasters have noticed with many of the long-lived supercells, the cycling of probability of tornado values related to cycling of mesocyclone strength.  This is giving forecasters good confidence in how to use the probability of tornado information in their decision making.

Figure 1.  KICT 0.5 degree storm relative velocity and ProbSevere All Hazards probability of tornado 2342 - 2358 UTC 01 May 2018.

-J. Sieglaff

NUCAPS/RAOB Comparison Over Dodge City, KS

The Dodge City, KS WFO launched a special 1900 UTC sounding, roughly the same time as a NUCAPS pass. Therefore, comparisons were made between the RAOB and a nearby NUCAPS. The NUCAPS profiles were not automatically modified, so user modification was needed. The surface temperature and dew point were both adjusted up roughly 5 degrees to match that of the radiosonde profile. The rest of the NUCAPS profile matched very well with the raob in terms of general shape. The 500 mb temperature in both profiles was -12C, while the 300 mb temperature was -43C in both. The NUCAPS profile captured drying between 700 mb and 500 mb, and moistening above. Upon modification of the NUCAPS surface conditions, the thermodynamic fields matched very well. ML CAPE was around 2000 j/kg in both, while SBCAPE was around 3060 in NUCAPS and 3200 in the raob. The LCL in the NUCAPS profile was around 5100 ft, while that in the raob was closer to 5500 ft.

On a similar note, a quick way to gauge the accuracy of the NUCAPS LCL in the absence of a radiosonde is to compare with cloud bases (if present) in surface observations. In this case, the DDC surface ob reported FEW at 5500 ft, matching that of the radiosonde.

Finally,

- Bucky