Sky poetry. Supercell near Luther, Oklahoma on March 25, 2015.

By the way, if you're interested in tornado chasing, you might enjoy my newsletter. It covers chasing tips, recent events, and tornado science. Delivered on the 1st Saturday of every month. Sign up here:.

Sources:. National Weather Service. An Overview of Environmental Conditions and Forecast Implications of the 3 May 1999 Tornado Outbreak (Roger Edwards & Rich Thompson). 10/10.

So what can we learn from this outbreak? Here are three thoughts:. 1. When the cap is weak, watch out for storms east of the dryline. 2. Look for subtle boundaries that could initiate storms. 3. Monitor wind shifts that can make wind shear stronger. 9/10.

Finally, a subtle "confluence boundary" initiated the first storms of the outbreak. This boundary was located well east of the dryline. Winds east of the boundary came from the southeast, increasing the wind direction shift with height (i.e., this increased shear). 8/10

The cap was pretty weak, too. Generally, this is unfavorable because you're likely to get a line or cluster of storms. But in this case, the lift was also weak. So, the storms were ideally-spaced to take advantage of the supercell ingredients. 7/10

So, then, what exactly made it unique?. Well, the deep-layer shear was strong but it was not exceptional. And many smaller events have had stronger low-level shear. Maybe less is more?. 6/10

And like other big days, mid-level lapse rates were steep -- over 8 degrees Celsius/km. Combined with deep, high-quality moisture, it created strong to extreme instability. This meant that air near the surface was primed to rise explosively -- like a volcanic eruption. 5/10

And like May 3rd, those events had rich and deep moisture. That day, dewpoints were in the mid-upper 60s. And the moisture layer was over 100 mb deep. It set the stage for warm downdrafts and strong instability. 4/10

But in some ways, those doom prophets weren't wrong. Like future setups, May 3rd had a prominent negatively-tilted trough. That is, the "dip" in the 500 mb height field sloped from northwest to southeast. 6 out of the Top 10 outbreaks have had this kind of trough. 3/10

When I began chasing, every big setup was supposed to be "the next May 3rd." . But it never happened. And people who made that call were made fun of. 2/10.

The May 3, 1999 tornado outbreak was a once-in-a-lifetime event. 10 tornadic storms, 69 tornadoes, 4 of which were violent (F4/5). The setup was so unique, you might dismiss it as a fluke. But it still offers important lessons for future forecasts. 🧵. 1/10

A High Risk was issued for a good part of Arkansas and Missouri. Needless to say, it verified!

The multiple-vortex phase of the powerful Pierce City, Missouri tornado of May 4, 2003. This was just one of several long-track, strong to violent tornadoes that formed that day.

This gnarly tornado tracked across rural parts of northern Oklahoma on May 1, 2008. Moisture was pretty iffy in the morning, with dewpoints only in the 50s. But a surge of >65F dewpoints arrived exactly when storms formed. A classic case of "just in time" moisture!

On November 16, 2015, I chased this massive tornado near Pampa, Texas. Officially, it was rated EF3. But I have little doubt it was capable of much stronger damage.

By the way, if you're interested in tornado chasing, you might enjoy my newsletter. The next one will dig into a science topic. Coming to inboxes on August 2nd!. Sign up here:.

McLean, Texas is another tornado magnet in the Plains. I’ve witnessed three tornadic storms near it. And the views aren't bad, either!. May 16, 2017

RT @wxkylegillett: The dusty Dickinson wedge tornado and double rainbow on June 19th. #NDwx

Rare anticyclonic tornado near Calumet, Oklahoma on May 29, 2004. It was part of a classic High Risk tornado outbreak over the central and southern Plains.