SPC First Moderate & High Risk Climatology

[Patrick Marsh]

Every year there seems to be a thread wanting everyone to guess when and where the first moderate or high risk will be issued. I went ahead and did a climatology of the first SPC moderate and high risks for the years 1990 - 2008. You can read more about my methods and see the figures by clicking here.

In summary, the first moderate risk (at any issuance time) occurs on 28 January in either NE TX and SW AR or southern AL. The first high risk occurs between the 27-28 March in SE AR and northern MS.

 

[Patrick Marsh]

I had a couple of offline requests for climatologies of moderate and high risks, so I've added them here.

Summary: Central OK "wins" for the greatest likelihood for moderate risks with an average of 30 per year. NE Kansas through WC Illinois "wins" for high risks with an average of 2 per year.

Please note that this climatology accumulates ALL outlooks for a given year. Meaning central OK shouldn't expect 30 days with moderate risks....merely 30 outlooks (1200 UTC, 1630 UTC, 2000 UTC, etc) for the year will have a moderate risk.

 

[Patrick Marsh]

Last post tonight, I promise.

I wanted to add that between 1990 and 2008 there were 3454 moderate risks and 243 high risks issued. The disclaimer above applies to these numbers.

 

[DWest]

You put quite a bit of work into preparing those graphics - to say the least they're interesting. The moderate risk data doesn't surprise me, but I am somewhat surprised the high risk outlooks probabilities don't also focus over the Oklahoma region as well.

 

[Jason Foster]

Cool....I often play with these type of stats....more for my own planning in the plains. Thanks Pat for running the numbers...I was really slow this year, focused on other priorities. Statistics are so fun.

 

[Patrick Marsh]

I've got plans to do things such as when a given outlook type is typically issued for a given year. That will be awhile though because crunching the numbers will take a while.

 

[Skip Talbot]

Thanks for creating these plots, Patrick. I find it interesting that OK, NE, and western KS didn't score higher in the high risk climatology. The OK panhandle is almost exempt. I can think of many good chases that occurred out there, and even some outbreaks, but I guess it goes to show that the SPC outlooks and best chase setups do not line up, and that most of the best chase days out there are slight risks. The high risk plot does show a striking pattern to areas I do regularly target on a chase, however. Probably close to a third of my chases have involved setting up or targeting the KS/NE/MO/IA corner, or the Jacksonville, IL area and points west toward the Mississippi. It seems like there is always triple point setting up in NE KS, with a warm front extending east across northern MO into IL. There must be some common patterns that setup across this region favoring tornado outbreaks (or at least forecast to favor tornado outbreaks). Despite the frequency with which I target this area, my best chases have been outside the high risk zones, with most of my tornado intercepts further west or north of the zone. This might be due to difficulty in chase high risk events (high shear/fast storm motions/too many storms) and probably the terrain difficulties along the MO and MS rivers. The western part of Illinois highlighted by the high risk zone does seem to be a tornado hotspot from my personal experience. I've had a couple intercepts there, and have heard of many more.

 

[Patrick Marsh]

And for completeness, I just finished the slight risk climatology. This will most likely be it for a while (unless someone else can come up with an idea) until I have a chance to code up the routine to determine when to expect the first issuance of a given risk category for each grid point.

 

[Connor McCrorey]

It would be interesting to see similar maps for SPC tornado probabilities. I have a feeling the "high risk" areas would be concentrated farther southwest if it specifically plotted 30% tornado probabilities or higher. Either way, nice work!

 

[mikegeukes]

Great work Patrick.
It would be interesting to see a map of all the slight, moderate and high risks combined.
Mike

 

[Derek Weston]

243 high risks over 18 years? Perhaps my knowledge of this issuance history is poor, but that seems like an awful lot. 13.5 high risks on average per year?

 

[Skip Talbot]

243 high risks over 18 years? Perhaps my knowledge of this issuance history is poor, but that seems like an awful lot. 13.5 high risks on average per year?

Well, the stats are per outlook not per day. So over 13 a year sounds reasonable to me if a high risk day has three or four outlooks that are high, making three or four high risk days.

 

[Jason Foster]

You know what would be outstandingly phenomenal, but likely way too much work...is have a GIS type system were you can have lots of 'layers' that include everything from visual sat., radar, metars, surface analysis, reports, and many more products overlaid with the various risk maps. Along with the graphic interface, perhaps a two panel screen that shows various graphs with data of the products being shown to help correlate the map/gis graphic data to the numerical data.

Sounds crazy, hard, time consuming though doesn't it.

 

[Patrick Marsh]

243 high risks over 18 years? Perhaps my knowledge of this issuance history is poor, but that seems like an awful lot. 13.5 high risks on average per year?

In the coming days I'll work on doing high risk days. This number should be a bit more intuitive.

 

[Patrick Marsh]

You know what would be outstandingly phenomenal, but likely way too much work...is have a GIS type system were you can have lots of 'layers' that include everything from visual sat., radar, metars, surface analysis, reports, and many more products overlaid with the various risk maps. Along with the graphic interface, perhaps a two panel screen that shows various graphs with data of the products being shown to help correlate the map/gis graphic data to the numerical data.

Sounds crazy, hard, time consuming though doesn't it.

You are right, it would be "crazy, hard, time consuming", but that's not the biggest hurdle to someone like me doing this. The problem will be finding computational resources for processing, storing, and serving up this large amount of data.