Andrew Lyons
banner
twisterkid34.bsky.social
Andrew Lyons
@twisterkid34.bsky.social
Forecaster. Storm and sky enthusiast. Occasional adult.

forecast.twodogswx.net

https://github.com/adlyons

For Pics &

Youtube.com/twisterkidmedia

For tornado videos

Bad opinions are my own.
Reposted by Andrew Lyons
I added a CIG1 for hail in the 20Z update today. I am just barely skirting in with those lines, I probably should’ve expanded further north a bit.

But, it looks like the environment had just enough juice to **potentially** produce 2” hail.

MESH is not a hail observation, but it’s something!
September 23, 2026 at 11:51 PM
Reposted by Andrew Lyons
VERIFI-CATION!
September 24, 2026 at 1:47 AM
Reposted by Andrew Lyons
Forgot to post this yesterday, but a decade ago @twisterkid34.bsky.social and I would talk about working at SPC one day. Pretty cool moment to train on the outlook update with him and issue a joint convective outlook.
September 23, 2026 at 6:11 PM
Reposted by Andrew Lyons
I’ve been trying to formulate opinions about this for a while. I still don’t know how to communicate what I am experiencing, thinking, & feeling. Lot’s of nuance.

The truth is LLM coding tools are too efficient & effective to be ignored anymore. Yes they’re flawed, but so is human intelligence.
September 11, 2026 at 10:57 PM
Im glad that agentic coding tools are making it easier for folks to realize ideas and make cool tools. However, the amount of slop with the exact same dark mode color scheme, layouts, and weird data visualization is low key infuriating.
August 21, 2026 at 8:07 PM
Some NWS people need to learn what rage and engagement bait are...
August 12, 2026 at 4:36 AM
Trying to post here more.

Some life updates.

Got married in June!

Went to Antigua 🇦🇬 in July!

Been a fun summer so far and ill have some professional news coming soon!
July 29, 2026 at 12:19 AM
Alot of discourse about the likely significant tornado in WI earlier. One thing I havent seen mentioned was the potential impact of a likely N-S boundary from the precceding storms. These were not well forecast and their impact is difficult to know but i think they played a role. (loop below)
July 28, 2026 at 12:35 AM
I do not know futbol as well as I know football, but I know when I see a spread offense winning on the perimeter with speed and passing.

Paraguay looks completely paralyzed in the middle of the pitch. Reacting late and flat footed to attacks and throughs from all sides.
June 13, 2026 at 2:00 AM
Reposted by Andrew Lyons
Today, I hit a pretty big life moment and got engaged to my best friend and favorite person in the world. Got so nervous that I accidentally placed myself behind the railing. Oh well 🙃😂. Celebrated with a lovely picnic in Central Park!
May 28, 2026 at 9:06 PM
Today looks fun, at my house.
April 25, 2026 at 4:50 PM
Reposted by Andrew Lyons
Today is SPC's first explicit forecast of significant tornadoes possible (CIG1) in a tornado probability less than 10%. Based on radar data and environment a significant tornado may have occurred in parts of that forecast area. SPC could not have explicitly forecast this kind of scenario previously.
March 6, 2026 at 4:54 AM
Reposted by Andrew Lyons
The NWS Storm Prediction Center's severe weather outlooks will be changing slightly next week.

Meteorologists Evan Bentley and Liz Leitman, our experts at SPC, provide an update about the new conditional intensity outlook.
February 27, 2026 at 8:42 PM
The way this snatch and grab went it may not even be real, which would just be peak trump. Non zero chance they convinced Maduro to go on his own with help.
January 3, 2026 at 1:03 PM
Shoutout to the No War president for starting a new war for "very legitimate" reasons.

Who could have possibly seen this coming!? /s
January 3, 2026 at 12:37 PM
My 3 year old ponderosa lemon tree made its first fruit and wow it was a chonker! 15 Oz! We turned it into a nice cocktail for new years.
January 1, 2026 at 2:58 AM
Reposted by Andrew Lyons
Maybe all of the people in leadership sending emails about how thankful they are for us showing up should do something about the fact we were all denied bonuses this year.
40+ days.

That was awful.

I dare someone to send one more “welcome back to work” email.

I BEEN HERE.
November 13, 2025 at 7:12 PM
Time lapse from my Lumix also turned out pretty well for minimal setup
November 12, 2025 at 4:28 PM
Last night still doesn't feel real. Visible aurora from 35° N over my backyard city lights. Insane!
November 12, 2025 at 4:10 PM
I was going to wait until the shutdown was over but the cat is out if the bag. My first journal article is now live!

Convective Mode Classification and Distribution of Contiguous United States Tornado Events from 2003–2023 in: Weather and Forecasting - Ahead of print share.google/M3ty4pVs5Tld...
Convective Mode Classification and Distribution of Contiguous United States Tornado Events from 2003–2023
Abstract Tornadoes and severe thunderstorms represent a significant threat to life and property in the United States annually. Approximately one thousand tornadoes, two-hundred being significant (F/EF-2+), are documented on average every year. Using archived radar and Storm Data storm report information, 21 912 tornado grid hours were manually analyzed for convective mode from the years 2003–2023. This dataset builds upon prior work by Smith et al. (2012) (hereafter S12) by more than doubling the sample size to produce a robust, multidecadal climatology of tornadoes by convective mode. Comparisons were made between tornado samples spanning 2003–2011 and 2012–2023 to assess changes in the frequency and spatial occurrence of tornadoes. Convective mode characterization consisted of a subjective analysis of WSR-88D imagery into three categories: 1) supercell, 2) quasi–linear convective system (QLCS), and 3) disorganized. Spatial climatologies of the different modes were performed and Kernel Density Estimate plots of events per decade were generated as in S12. The highest climatological frequency for tornadic supercells is reaffirmed to extend from KS and OK east-southeastward to MS and AL, while QLCS tornadoes are more frequent from the northern Gulf Coast states into the lower Ohio River Valley. Comparing the new sample to the original, QLCS tornado relative frequency increased by over 100%. Substantial variation in QLCS tornado occurrence may be due to non-meteorological factors such as dual polarization radar and different observing/reporting practices.
share.google
November 5, 2025 at 3:18 PM
Im 7 days behind because mids. Heres 4 #shutdown sammys.
October 26, 2025 at 3:21 AM
#shutdown sammy and #restorefunding ralphie for days 16 17 and 18 of ??

Im on mids so my captions are creatively limited but the dogs are cute.
October 18, 2025 at 7:29 AM
#shutdown sammy day 15 of ?? Sammy is perplexed by the lack of ball and a functional government. But mostly ball.
October 16, 2025 at 3:48 AM
#shutdown sammy and #restorefunding ralphie days 13 and 14 of ??

I just like this one because of the tounges.
October 15, 2025 at 1:38 AM
#shutdown sammy #restorefunding ralphie day 11 and 12 of ?? Pups in the pool. Where's our pool of money?
October 13, 2025 at 3:04 AM