Kevin Gray (@KevinGrayWx) header
Kevin Gray (@KevinGrayWx) profile picture

PhD in atmospheric science. Postdoc at UIUC. Current research: Extreme rainfall in a warmer climate, simulated supercells, and streamwise vorticity currents.

Joined July 2016
Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)Photo by Kevin Gray (@KevinGrayWx)
As a last promotion for the new paper (doi.org/10.1175/MWR-D-25-019…), here's a peak at one of our figures. Plotted is mean streamwise vorticity below 2 km (shaded and blue contour), reflectivity (black), and surface convergence (red).
1
1
7
408
The panels within the green box in the second figure depict SVCs from simulations using a semislip lower boundary. All the other panels are SVCs from freeslip simulations. See what differences you notice! We dive deeper in the paper by using cross-sections and trajectories.
1
61
The last of my PhD work with @VORTEXJeff titled "Characteristics of Streamwise Vorticity Currents in Simulated Supercell Thunderstorms Using Free-slip and Semi-slip Lower Boundary Conditions" has been accepted for publication in @ametsoc MWR!
2
1
21
986
We found that surface drag has little impact on SVCs along FFCBs since they develop farther above the surface. SVCs along LFCBs develop closer to the surface and exhibit decreases in streamwise vorticity and increases in crosswise vorticity as trajectories approach the surface.
2
3
137
Our article is now available in Monthly Weather Review here: doi.org/10.1175/MWR-D-25-019…
1
5
8
767
If most supercells are capable of producing sufficient sub-tornadic vorticity near the surface to be stretched (Coffer and Parker 2018), then persistent convergence beneath the low-level mesocyclone is the name of the game for increasing the probability of tornadogenesis.
In 109 supercells, tornadic storms tended to have stronger, more organized rear-flank winds directed toward the low-level mesocyclone. Buoyancy differences were small, suggesting that outflow organization mattered more than whether it was warmer or colder. #AMSMadisonSummit
2
24
1,830
SVCs along FFCBs may exhibit less streamwise vort than SVCs along LFCBs, but they may be less susceptible to the effects of surface drag. We also note other interesting storm behaviors between our free-slip and semi-slip simulations. Early online release link will be shared soon!
4
78
📢Researchers using CM1...I created a web-app to interactively generate input_sounding files. You can upload your own SHARPpy format, resample to desired dz, GUI edit, table edit, and save for use in CM1. Happy storm creation! atlas.niu.edu/cm1gen/
10
35
2
163
10,776
Just ran HRRRcast for the first time. Tested a 6 hour forecast, each hour time step took ~70 s. Our research team @CliMAS_Illinois will likely use a HRRRcast ensemble to produce probabilistic forecasts of heavy rain. These forecasts can then help with flood prediction.
8
642
I believe this is where the radar that detected the first hook echo was once located. It’s at Willard airport and you can see it from hole 14 of university of Illinois golf course (the blue course) if you’re ever in Champaign!
1
3
7
1,131
Play suspended at the Philadelphia Eagles stadium, rightfully so, for lightning in the area. But they broadcasted a severe thunderstorm warning message on the stadium PA, which appears to be inaccurate?
3
396
We're working on using @nvidia 's CorrDiff model to help with uncertainty in rainfall forecasts. The model is trained to take in GEFS data and generate a HRRR-like output. GEFS has 31 members and CorrDiff can generate several samples, giving many possible rainfall solutions.
1
7
321
The beauty here is we’re just using forecasts that already exist and downscaling them in a few minutes using CorrDiff. The images above are forecasts for tomorrow at 18z. I generated 4 samples, two look realistic and the other two forecast no significant precipitation over IL.
1
1
128
Radar mosaic at 18z (Aug 12) for verification.
64
Kevin Gray retweeted
For years, I've wanted a way to get a text when friends or family across the country were under critical NWS warnings—so I built it. Meet nausicawx.com. Tornadoes, severe storms, flash floods, freeze warnings, and more. Get instant SMS alerts for multiple locations.
1
1
4
283
Targeted Lawton, OK, for the last day of our trip. We’ll see if one of these can take off out ahead of the cold front farther west.
3
116
Wind incoming 💨
14
549
Drove 13 hours on our first day to see a nice cell near Burlington, CO.
2
10
322
Dropped south to the next cell, saw a tornado, then we were treated to the sunset just below cloud base. Great start to the trip!
6
133